Focusing issue with the 35mmf 1.4L II

[Eldar]

I got mine about a week ago. Optically it seems to be everything I was hoping for, BUT! autofocus is inconsistent. I have tried it on both 5DSR and 1DX, with the same result. Most of the time it is fine, with the manually verified AFMA setting I got. I had to do it manually, since FoCal was unable to set an AFMA value. It can front focus on one shot and back focus on the next. It can also be totally off, while the camera(s) still report focus lock.

To me it looks like the Sigma Art story all over again. I´ll deliver it to CPS tomorrow, so we´ll see what they find out.

Has anyone else experienced the same?

 

[Viggo]

That's the first I've heard of an inconsistent L, and it's the worst possible lens for it to happen to. Wow... I've had lots of "foCal has tried 5 times.. Blabla", but it's always the setup, except with the Art lenses. But this is simply horrible news. There shouldn't be any inconsistency issues with a new L.

 

[gjones5252]

Had mine for about three weeks and all focusing errors were user error.
I would actually say the opposite it spot on all the time.

 

[Eldar]

I am not basing my conclusions on individual, regular shots. I am using a LensAlign rig for manual AFMA and it was when using this it became apparent that the focus is inconsistent. But if nobody else has the same experience, I´m sure CPS will fix it. Fingers crossed!

 

[Rahul]

Mine is tack sharp on the 5d3 as well as the 6D - shot about 500-600 frames in the last week with this lens. I've not had the time for AFMA yet (and I don't think that my copy needs AFMA). I am yet to get an OOF shot that I've not been able to diagnose as "user error".

 

[TeT]

Its the lens. You just bought it, was it local? If so... take it back for an exchange and save yourself the week plus of no lens. It is obviously faulty so there should be no exchange issue...

 

[zim]

Its the lens. You just bought it, was it local? If so... take it back for an exchange and save yourself the week plus of no lens. It is obviously faulty so there should be no exchange issue...

+1
Seen enough of Elder's work to not believe user error
Actually I wonder if this is what's happened at Dx0

 

[TeT]

No, from what I can gather DxO liked all aspects of the Canon lens except that it had to be attached to a Canon Camera for the test... and / or that it said Canon on it...

 

[eml58]

Hi Eldar
Seems you may not be the only one having this issue, won't help but you could take a look here.

http://www.fredmiranda.com/reviews/showproduct.php?product=369

From what I'm seeing there are very few people having any issues, seems though you may have received one that does, so perhaps an exchange may fix the issue.

I have the older 351.4L and may buy this one as well, but in the meantime I'm waiting expectantly for the Otus 28, I'm in Europe with the Family for 5 weeks this Xmas New Year so the 28 would have been welcome, still, the 21 is not so bad either.

Hope to hear your issue gets sorted positively.

 

[StudentOfLight]

Hi Eldar
Seems you may not be the only one having this issue, won't help but you could take a look here.

http://www.fredmiranda.com/reviews/showproduct.php?product=369

From what I'm seeing there are very few people having any issues, seems though you may have received one that does, so perhaps an exchange may fix the issue.

I have the older 351.4L and may buy this one as well, but in the meantime I'm waiting expectantly for the Otus 28, I'm in Europe with the Family for 5 weeks this Xmas New Year so the 28 would have been welcome, still, the 21 is not so bad either.

Hope to hear your issue gets sorted positively.

I'm confused, the link you posted pertains to the 24L II, not the new 35L II.

 

[ben805]

most likely you have a defective copy. I shot a wedding over the weekend with over 800 shots with this lens, and all of them are spot on. Only missed a few due to user error when i had my 5D3 on servo mode with the focus point on the wrong spot.

 

[Eldar]

I have now had a couple of rounds with CPS and I am not a very happy man. I had the lens in for repair and got it back untouched. There is nothing wrong with this lens. Then I go home, I set up LensAlign rig, 1.75m from the lens, one shot, single focus point, the camera is set to minimum focusing distance prior to each shot. Below you see the three first shots I made. I have not attempted to find the extremes, because it is obvious to me that this lens is inconsistent. Check the difference between the first and the third.

According to CPS, they have a more advanced setup (I´m sure they have), which proves that the lens is OK and claims that my test is invalid and the whole thing is user error.

So; Have a look and tell me what is wrong with my setup and why I should accept this as proper performance.

For those of you with more insight;
What is the DoF for a 35/1.4 lens at 1.75m?
What is the accepted focus deviation on a lens like this?

 

[YuengLinger]

Are you still within the return window?

The only unpleasant experience I've had with CPS was with version I of this lens. Same--they just said ''in spec,'' nothing to fix.

I sold it and got a good Sigma. (But could not get a good 50 Art.)

Really hope you get it sorted!!!

 

[kaihp]

For those of you with more insight;
What is the DoF for a 35/1.4 lens at 1.75m?

Eldar, according to the DoF Calc app on my smartphone, the DoF should be 19.7cm at 170cm focus distance*, in the bracket 160.7cm - 180.4cm.

http://www.dofmaster.com/dofjs.html yields a 20.9cm DoF, bracketing 165.2cm-186.1cm.

I'm really sorry to hear that you have these issues - are you some kind of lens problem vortex?

*) The app won't allow me to pick 175cm, only 170cm and 180cm.

 

[Viggo]

Here's a stupid question, is it the same with the 1dx and 5d?

 

[bigdaddy]

http://www.dofmaster.com/dofjs.html yields a 20.9cm DoF, bracketing 165.2cm-186.1cm.
[/quote]

That calculation by dofmaster seems to be somewhat off. The means a range of 20 cm (~8 in) section of the photo should be in focus. That does not look to be the case with the target Eldar is using. I would have expected a much smaller focus range, like 2 cm or something, at that distance and at 1.4.

@Eldar

Maybe its the camera? Can you reproduce the inconsistencies on different bodies?

Have something so expensive not work, and then be told everything is good is really a pain. I hope you find a solution soon.

Take care,

bigdaddy

 

[bigdaddy]

Here's a stupid question, is it the same with the 1dx and 5d?

Ah, you beat me by a few seconds.

 

[Eldar]

I did of course suspect the camera, so I repeated the test with my 1DX. This is my work horse, which, apart from a few very difficult tracking situations, have never let me down. Below are the 3 first images I shot. They are so far off that I do not see the point in getting more samples.

 

[kaihp]

That calculation by dofmaster seems to be somewhat off. The means a range of 20 cm (~8 in) section of the photo should be in focus. That does not look to be the case with the target Eldar is using. I would have expected a much smaller focus range, like 2 cm or something, at that distance and at 1.4.

Your scepticism is appreciated - it did seem high to me too. For the heck of it, I checked it against http://cpn.canon-europe.com/content/education/technical/depth_of_field_calculator.do : 165.3cm - 185.9cm = 20.6cm
http://www.cambridgeincolour.com/tutorials/dof-calculator.htm : 164.67cm - 186.71cm = 22.04cm
http://dofsimulator.net/en/ : 165cm - 186cm = ~20.7cm

They all give a result in the 20-22cm range, so the formulaes agree and we can disspell that as a source of error.
I think the "problem" comes back to what the definition of Near & Far "distance of acceptable sharpness" is. In this case, we're probably significantly more picky than what the formulaes assume for the CoC.

 

[rfdesigner]

That calculation by dofmaster seems to be somewhat off. The means a range of 20 cm (~8 in) section of the photo should be in focus. That does not look to be the case with the target Eldar is using. I would have expected a much smaller focus range, like 2 cm or something, at that distance and at 1.4.

Your scepticism is appreciated - it did seem high to me too. For the heck of it, I checked it against http://cpn.canon-europe.com/content/education/technical/depth_of_field_calculator.do : 165.3cm - 185.9cm = 20.6cm
http://www.cambridgeincolour.com/tutorials/dof-calculator.htm : 164.67cm - 186.71cm = 22.04cm
http://dofsimulator.net/en/ : 165cm - 186cm = ~20.7cm

They all give a result in the 20-22cm range, so the formulaes agree and we can disspell that as a source of error.
I think the "problem" comes back to what the definition of Near & Far "distance of acceptable sharpness" is. In this case, we're probably significantly more picky than what the formulaes assume for the CoC.

As we have 100% samples to look at we're really looking at the equivelent of 20/20 vision staring intently at an A3 print. Stuff those numbers into Cambridge in Colours DOF (using the advnaced tab) and you get something in the order of 4cm DOF.. Now I seem to remember that the Mk1 Canon EF system (5DII/7D/6D and before) merely guarenteed to get the focus point within the DOF. But the 1DX, 5DIII, 7DII etc AF sytems have superior focus accuracy on the double cross centre spots (see Rogers discourse on Lensrentals)

My understanding is the DOF variation should be about 1/2 the DOF range with these newer cameras... I think you also need a newer lens (70-200F2.8 IS II and later), the 35LII certainly counts.

So assuming you're only using centre spot, then the 0 point should always be in focus.. no matter what.