Focus issue: 300/4.0L (non-IS)

All my lenses work great on my R (and 5D series bodies too), except I finally took out the 300/4L, the old non-IS version, and it won't give me good focus on the R.

I did some testing, and here's what I found:

Autofocus - no matter the spot, one shot of course, and its pretty much off. I think I said backfocus issue on another web forum, but its actually front focus.
Manual focus with peaking - off like AF
Manual focus with triangles - off like AF

Manual focus with 10x - spot on, as good as with my 5d series cameras. The 300/4L doesn't need "correction" with any of those bodies, btw.

Should also add that its the same with the old "regular" 1.4x as well, except more pronounced.

So... I contacted CUSA, and we'll wait and see what they say.
 

JPAZ

If only I knew what I was doing.....
CR Pro
Sep 8, 2012
1,164
641
Southwest USA
Since the "focus sensor" and the "image sensor" are the same on a mirrorless, and since the camera seems fine with other lenses, it would have to be an issue with the 300/4L. Does the lens work well with your 5D bodies? If so, then I'd have to agree it is a "communication" issue. If not, I'd wonder about things like elements shifting out of alignment, etc.
 
Upvote 0
I finally heard back from Canon, and they really flubbed the response....

Thank you for contacting Canon product support. I understand that you have questions regarding your EOS R and EF 400mm f4L USM lens. I can address this for you.

I would suggest that you try cleaning the lens contacts on the
EF 400mm lens, but you mentioned the lens works fine with other cameras. This would lead me to believe that this may be something else. I have not heard of this combination or problem. I will forward this information to our engineers and see what they have to say. I will keep you informed on any findings. In the meantime, continue to check the Firmware options on the EOS R page to see if there are any updates posted.

Please let us know if you have any other questions or concerns.

Thank you for choosing Canon.


I tried cleaning the contacts... same result. With a 5D3, no focus correction, the focus is as close to "spot on" as I could ask for. Same when using my old 5D2. That's using the center focusing point only (and not the "fine spot" the 5D3 offers).

With the R, it focuses a whole lot closer, maybe about 6-8 feet at 75 feet distance. In fact, if you manually focus with the 10x magnifier, then have the camera auto focus, you can see the image "pop out of focus" in the EVF.

The 300/4L (non-IS) is such a stellar performer, and has been ever since I got it to go with EOS5 (film) bodies, many years ago. Its basically like shooting the "sharpness of a good 50mm lens", but its 300mm. Just crisp and sparkly with decent bokeh too.

Next test is to run lots of non-center focus points on the R. Fair weather today, so it shouldn't take too long.
 
Upvote 0
And that didn't take too long at all.

Tried focusing point on the R at

centered vertically, 1/3 out to the left edge, then extreme out to the left edge
centered horizontally, 1/3 out to the top, then extreme out to the top
1/3 x 1/3 left and upper, then extreme left and upper spots

Didn't make any difference, it all focuses a good deal closer.

Let me also say... I'm focusing on a lifted asphalt roof tile, on my back neighbors roof. The tile is about 1/2 way up the roof. Google imaging suggests I was off by a bit in my distance measurement, with the distance being 99.75 feet, and the focus being offset closer to the camera by 7 feet.

If its clear, I'll try the 3/4 full moon tonite, and see how it does as infinity.
 
Upvote 0
Mar 25, 2011
16,848
1,835
And that didn't take too long at all.

Tried focusing point on the R at

centered vertically, 1/3 out to the left edge, then extreme out to the left edge
centered horizontally, 1/3 out to the top, then extreme out to the top
1/3 x 1/3 left and upper, then extreme left and upper spots

Didn't make any difference, it all focuses a good deal closer.

Let me also say... I'm focusing on a lifted asphalt roof tile, on my back neighbors roof. The tile is about 1/2 way up the roof. Google imaging suggests I was off by a bit in my distance measurement, with the distance being 99.75 feet, and the focus being offset closer to the camera by 7 feet.

If its clear, I'll try the 3/4 full moon tonite, and see how it does as infinity.
Being a older lens, its entirely possible that the camera firmware does not play well with it, however, issues with firmware are usually not so obvious. You are likely going to end up sending the lens in for testing on their standard calibrated EOS R body. If its OK, then your body has a issue. Since the lens works well with other camera bodies, that points to a body issue or compatibility issue, but I'd have the lens tested against the Canon test body.
 
Upvote 0
Well, I know the "reviews" of the new(er) 300/4L IS are so glowing, but there's something to be said for 7 elements in 8 groups, versus 15 elements in 11 groups.....

I'm hoping the sparkle, the vitality of the 300/4L IS is up to the older lens. I never had an issue with image stabilization (or lack thereof), as I'm used to shooting long lenses (can you say 350 Tele-Tessar and a Mutar 2x on 6x6).

At any rate, I pulled the trigger on a 300/4L IS from the Canon Store, refurb, with the extra 15% off. Lets see if that lens is up to the task on the R (and the 5D2 and 5D3 bodies)

I've got a friend with a 7D who'll love the gift of an older 300mm L lens.... so no biggie there.
 
Upvote 0
I need to add this, and ask anyone with a host of lenses to see how their entire brood behaves on their own EOS R....

EF 20mm/2.8 USM slightly front focusing, nothing too bad, DOF at 2.8 takes care of it.
EF 100/2.8 USM (2nd, current non-L version), slightly back focuses when focusing under 1 meter or so. At distance, its fine, and I'm usually manually focusing for macro anyway.
EF 85/1.8 USM - a little back focusing at very close distance, works great at head/shoulder and distance
EF 200/2.8 L - a little back focusing at very close distance, same as the others... but ok at distance and medium to close distances, like about 10 feet.

Those above, are really "ok" for my uses, as I'd have to be very close and wide open for a problem to show, and even then, its pixel peeping detectable, would never see it on a print, or on a screen unless peeping.

EF 28/1.8 USM - right on
EF 50/1.4 USM - right on
EF 16-35/2.8ii L - right on at any focal length I've tried
EF 24-105/4 L - right on at any focal length I've tried


These two surprise me
EF 28-80 from an old film cheapie - right on
EF 80-200 from the same old film cheapie kit - right on

The point of this last post, is that its becoming more evident to me that the dual pixel magic isn't really all that magical. Yes I get the concept of the image sensor being the focus sensor, but I also get the reality that the information needs processing, and that the processing may be deficient, relying on generalities. I would be surprised if Canon didn't come out with a FA feature on some future R series body. I wonder why there is only 1 non-L lens in the current or foretold / projected RF lens lineup?
 
Upvote 0
Mar 25, 2011
16,848
1,835
I need to add this, and ask anyone with a host of lenses to see how their entire brood behaves on their own EOS R....

EF 20mm/2.8 USM slightly front focusing, nothing too bad, DOF at 2.8 takes care of it.
EF 100/2.8 USM (2nd, current non-L version), slightly back focuses when focusing under 1 meter or so. At distance, its fine, and I'm usually manually focusing for macro anyway.
EF 85/1.8 USM - a little back focusing at very close distance, works great at head/shoulder and distance
EF 200/2.8 L - a little back focusing at very close distance, same as the others... but ok at distance and medium to close distances, like about 10 feet.

Those above, are really "ok" for my uses, as I'd have to be very close and wide open for a problem to show, and even then, its pixel peeping detectable, would never see it on a print, or on a screen unless peeping.

EF 28/1.8 USM - right on
EF 50/1.4 USM - right on
EF 16-35/2.8ii L - right on at any focal length I've tried
EF 24-105/4 L - right on at any focal length I've tried


These two surprise me
EF 28-80 from an old film cheapie - right on
EF 80-200 from the same old film cheapie kit - right on

The point of this last post, is that its becoming more evident to me that the dual pixel magic isn't really all that magical. Yes I get the concept of the image sensor being the focus sensor, but I also get the reality that the information needs processing, and that the processing may be deficient, relying on generalities. I would be surprised if Canon didn't come out with a FA feature on some future R series body. I wonder why there is only 1 non-L lens in the current or foretold / projected RF lens lineup?
Phase detect does not check to see if the lens goes to the requested point, so if the lens itself has a error, the focus will be inaccurate. I had believed that the Canon DPAF implementation had a feedback loop that checked and adjusted the focus, but I can't find anything verifying that.

Then, there is the issue with focus shift. Some have said that their R stops down to focus with certain lenses, then opens for the exposure. A lens with focus shift will experience the problem with those lenses.

I have a 50mm Macro on my R sitting next to me, it does not stop down when focusing and metering, so if the lens had focus shift, it would shift focus when shooting at a small aperture.
 
Upvote 0
Well, an interesting development. Very interesting.

The new EF 300/4L IS came in today and it focused perfectly. After a bit of testing a pixel peeping, I decided that the new lens was a keeper. So got out the older 300/4L USM (non-IS) and took the filter off it.... and said... jeeze I better check this out.

As Gomer Pyle sez "Shazam Shazam!". After testing great, with no filter, the IS version front focused just like the older lens. So... scientific me.... I tried the non-IS that had (formerly) front focused, except with no filter. The non-IS focused perfectly with no filter.

I've been using Hoya HMC Sky 1B filters on all my lenses since... jeeze... as long as my seventh decade of life can remember. Never had an issue until... good ol' magic dual-pixel EOS R came along.

So, I will probably be sending the newer IS version back - remember I don't have a problem with the non-IS, so there's no need.

I genuinely feel rather stupid over this one... but I still don't know how the lens works great on everything I've had to date, with a filter.
 
Upvote 0
Ha, very interesting as you say. On a related point, I have also always used front skylight filters, as much for protection in the rough and tumble of shooting as for their UV cutting qualities. It has never previously occurred to me to try AFMA with and without skylights on, so I wonder if that makes any notable difference (generally, not just thinking of mirrorless).

Anyone ever done any side-by-side comparisons on this? I'm busy currently, but might give it a try out sometime in the next few weeks if I get a chance.

Stoical.
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,342
22,517
Well, an interesting development. Very interesting.

The new EF 300/4L IS came in today and it focused perfectly. After a bit of testing a pixel peeping, I decided that the new lens was a keeper. So got out the older 300/4L USM (non-IS) and took the filter off it.... and said... jeeze I better check this out.

As Gomer Pyle sez "Shazam Shazam!". After testing great, with no filter, the IS version front focused just like the older lens. So... scientific me.... I tried the non-IS that had (formerly) front focused, except with no filter. The non-IS focused perfectly with no filter.

I've been using Hoya HMC Sky 1B filters on all my lenses since... jeeze... as long as my seventh decade of life can remember. Never had an issue until... good ol' magic dual-pixel EOS R came along.

So, I will probably be sending the newer IS version back - remember I don't have a problem with the non-IS, so there's no need.

I genuinely feel rather stupid over this one... but I still don't know how the lens works great on everything I've had to date, with a filter.
The seller, unfortunately, will be the one who ends up paying for this having paid postage and having a used lens to sell at a discount.
 
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,342
22,517
Ha, very interesting as you say. On a related point, I have also always used front skylight filters, as much for protection in the rough and tumble of shooting as for their UV cutting qualities. It has never previously occurred to me to try AFMA with and without skylights on, so I wonder if that makes any notable difference (generally, not just thinking of mirrorless).

Anyone ever done any side-by-side comparisons on this? I'm busy currently, but might give it a try out sometime in the next few weeks if I get a chance.

Stoical.
It’s been reported in the past that 400mm lenses can have problems with filters and so I have always AFMAd mine with and without filters. In no cases have I found any differences.
 
Upvote 0
UPDATE:
I remembered that I have 77mm filter on my 24-105L, also a Hoya HMC Sky 1B
So, I tried that filter on the non-IS 300/4 lens, and it works perfectly, focus is well within DOF wide open.
Tried the same filter on the IS 300/4 lens, and it works perfectly, focus is well within DOF wide open.

Now I got curious, and tried the "suspect" filter on the 24-105L ... and I can't see any issue. Looking at the two filters, I dunno... they look identical, but obviously one is not the same as the other. Go figure. Also tried 82A filter, and it worked great too.

Getting very curious, I had to go back to my 5D3 and try the combination. Still focuses spot on, with "zero correction".

End conclusion - something about THAT filter, and Dual Pixel focus, or maybe just my R body (since I only have one and can't A/B compare), causes focus errors. I've had a short conversation with the folks at LensRentals (great guys btw). They thought it was interesting, and good to know.
 
Upvote 0