Lasse Eriksson wrote:
If the focus is different with center point and other AF points, it's a camera issue.
Camera + lens issue. But not a camera issue. My camera plays nicely with nikon fast primes. It is a nikon thing, it seems canon gets along alot nicer than nikon with sigma.
I was just asking about the dock, not looking to get into a debate. If the dock did have the capability to cover me with the issue I'd consider purchasing the lens again. But I guess not
Well, I'll leave the center, left and right point argument alone for now.
But I did play with the Sigma dock today and found it superb. I took a 35mm f/1.4 that was impossible to MFA on my camera (it was fine at near distances, backfocused badly at long distances) plugged it in and 10 minutes later it was behaving perfectly at all focusing distances with adjustment of 0,0,0, -8.
Svassh - simple, just back up! . Seriously though, just get a sufficiently large target and as far away as you reasonably can.
Roger- seeing how well this works, what implications does this have for your rental business? Will you start adjusting them before they are rented? Also, will you have to check them upon return to be sure the settings are still as you put them? Sounds like an excellent opportunity to offer more accurate lenses, but at a ridiculous cost of time. Just wondering out loud here...
RCicala wrote:
Well, I'll leave the center, left and right point argument alone for now.
But I did play with the Sigma dock today and found it superb. I took a 35mm f/1.4 that was impossible to MFA on my camera (it was fine at near distances, backfocused badly at long distances) plugged it in and 10 minutes later it was behaving perfectly at all focusing distances with adjustment of 0,0,0, -8.
That is cool. That is very cool.
Roger: did you find, as others have, that the MFA values on the lens = MFA values on the camera body? I'm assuming yes, since you didn't mention otherwise.
I don't know how useful this is but I have come up with refined distances for the Sigma 35mm AFMA fine tuning.
The method I used is as follows:
I set focus to manual on the lens.
I adjusted the manual focus ring until the distance scale on the lens matched the software illustration exactly.
I then moved the camera until I achieved optimal focus (with 10x magnified live view)
I then measured the distance from the target to the camera focal plane (not the lens front).
Here are the distances I obtained.
12.25"
14.75"
26.25" (possible discrepancy here because diagram does not match the distance scale exactly)
~80 feet (for infinity, had to do this one outside and I lost the light, will refine tomorrow)
Perhaps others can repeat the test and confirm the measurements. Since AFMA adjustments vary with distance it would be helpful if everyone used the same set of distances so we can compare results.
Has anyone tried to move an AFMAed lens to a different body (5DM3 to 7D for example) to see if tuning is accurate on a different body (assuming the 2nd body is properly adjusted for AFMA)?
I ask because I recall getting significantly different AFMA results when I moved my Canon lenses between bodies (the differences were not just by a fixed offset). Could be that my calibration method was the source of the error but I have since sold my 7D and can't do the tests again.
austin.grant wrote:
Roger- seeing how well this works, what implications does this have for your rental business? Will you start adjusting them before they are rented? Also, will you have to check them upon return to be sure the settings are still as you put them? Sounds like an excellent opportunity to offer more accurate lenses, but at a ridiculous cost of time. Just wondering out loud here...
I'm not sure yet. We need to confirm that a selective adjustment like the one I did today is going to be consistent between bodies. If that's the case we'll leave the Sigma lens I did today (0,0,0,-8) like that from the doc and then use camera MFA to fine tune, so if it was going to need 4,4,4,-4 on the next body, we'd leave the lens alone and let the renter shift the entire curve using camera MFA.
But I have a feeling that it won't be that simple. Even if this particular lens copy has the same curve on all Canon 6Ds, I doubt it will on a 5DIII or a T4i. Honestly, I'll be a bit surprised if it behaves the same on all 6Ds.
So probably we will zero them out and rent docks with lenses at a very nominal price to those who want to do their own adjustments. I'll need to confirm Sigma's software agreement doesn't require a purchased dock for download, though.
p.2 #10 · Sigma USB Dock calibration data for 35mm
Photonic wrote:
Has anyone tried to move an AFMAed lens to a different body (5DM3 to 7D for example) to see if tuning is accurate on a different body (assuming the 2nd body is properly adjusted for AFMA)?
I ask because I recall getting significantly different AFMA results when I moved my Canon lenses between bodies (the differences were not just by a fixed offset). Could be that my calibration method was the source of the error but I have since sold my 7D and can't do the tests again.
I think it's certainly going to vary with different types of bodies. It almost has to since they will (in most cases) have different AF systems. I think it's possible the curve might be the same for same camera AF systems though, (like between all 5DIII or all 7Ds). I'll try to check this on several bodies tomorrow.
p.2 #11 · Sigma USB Dock calibration data for 35mm
Photonic - good information on the test distances, and a very good question on the consistency across different bodies. Roger touches on the same idea, and I think that's the next step in evaluating this system. (Not that we've at all exhausted the original question of "Does it work WELL?"
Also, although the 50X focal length is generally recommended as a minimum distance for MFA, I think that going significantly beyond it shouldn't give significantly different results. Here's the graph from the Focal documentation:
p.2 #12 · Sigma USB Dock calibration data for 35mm
I must confess that I do not understand why Sigma picked these four distances but as someone else noted they must have had a good reason. Following that logic, I decided to set my lens exactly as shown in their software.
You pose a good question about the difference between 50x focal and infinity. I will check tomorrow when I have sunlight. Even with a mirror in my basement I can only do 60 feet indoors.
p.2 #14 · Sigma USB Dock calibration data for 35mm
Photonic wrote:
I must confess that I do not understand why Sigma picked these four distances.
In looking at the distance scale, those seem to be the distances you cover with 25% increments of rotation of the focusing ring.
Or looked at another way, each of the distances is about 25% of the motor/AF systems range.
It's quite different on the 120-300 and I'm sure will be for other lenses, but a wide-aperture close focusing 35mm lens needs a lot of precision for close focusing with that narrow depth of field, less so at further distances, so much of the focusing range is going to be up close. Or at least that's what the designers chose for this lens.
p.2 #16 · Sigma USB Dock calibration data for 35mm
I just spent a few hours characterizing my new Sigma 35mm using Dot Tune on a Canon 5DM3.
First a word about Dot Tune. I am extremely impressed with this method. I think of it as a stud finder. You find the left edge of the stud and then the right. You set your value right in the middle. If one edge is off the scale then you use your knowledge of the width to find the middle. I did all testing by listening to the beep alone (Beep Tune anyone?). I set 15 consistent and immediate beeps as my threshold for a valid sample. Any delay or missed beep was a disqualifier.
I tested 6 distances. I ran the test 3 times for each distance. The width in every case was either 14 or 15. The width was so consistent, in fact, that I would be confident finding only the left edge and adding half the width for center in the future.
The edges were always within one count of each other for each distance. This level of repeatability certainly inspires confidence. Once tuned for a distance I checked AF against live view focus and found the results to be spot on.
Side notes: I used the modelling lamps from two Elinchrom strobes for consistent illumination throughout. I found the results to be VERY sensitive to focus distance. I used a Bosch laser distance measuring device to confirm all distances. All distances are from camera focal plane.
12.25" +4
14.75" +6
26.25" +17
69.00" +15 (50x FL)
138.0" +11 (100x FL)
276.0" 0 (200x FL)
infinity (need to wait for sunrise to do test outdoors)
I was not expecting these results (especially at the longer distances). In the coming days I plan to go back and fill in the intermediate distances so I can plot the lens performance over the entire focus distance range.
p.2 #17 · Sigma USB Dock calibration data for 35mm
Finally got mine dialed in, the issue I was having was I would set the 4 lens MFA to 0 and the body to 0. Then with Focal I would do the 12" distance first and adjust the MFA using the dock, set the body back to 0 and then Focal it again to validate the setting was still 0. When I started in on the next distance of 16" the lens was set to 0 and Focal would be all over the place when I set the body to the various MFA amounts. I could never get it dialed in. What I ended up doing was dialing in the 12" distance then using that to set all 4 lens MFA amounts to that value. Then when I did the 16" for example Focal would say it needed -1 so I would subtract 1 from the current lens setting for 16". Long story short it appears to be quite sharp and my lens MFA is as follows:
p.2 #18 · Sigma USB Dock calibration data for 35mm
I have updated numbers for my lens
12.25" +4
14.75" +6
26.25" +17
69.00" +15 (50x FL)
138.0" +11 (100x FL)
276.0" +11 (200x FL) - corrected from first set, had to use larger target pattern at this distance
infinity +11 (I am still not confident in this result, too much variation test to test)
The longer distances are more consistent with what I would have expected. I also checked intermediate distances and they fell nicely between adjacent values.
p.2 #19 · Sigma USB Dock calibration data for 35mm
MikeW wrote:
because the focus fine tuning is different from left, centre & right. So unless you always shoot with only one focus point you are always front or back focused. For the nikon mount this issue has shown up, I unfortunately got landed with one. Someone has sent the body & lens to sigma & it came back perfect. I wasn't prepared to send in my D4, so I returned the lens.
I was interested in if the dock could fix the problem but it can't so I'll consider a nikon one. Or switch to canon
p.2 #20 · Sigma USB Dock calibration data for 35mm
Okay, so I've had my Sigma 35 calibrated and in use for several days now, and my preliminary results are not encouraging. Multiple Dot Tune and Focal tests show that the lens is MFA'd to spec, but the focus consistency on my copy is abysmal. I'll get out today or tomorrow and run some focus reliability tests with Focal to get some actual numbers, but rest assured it's bad.
I'm going to give it a couple more days of tweaking before sending it back, although I don't see how it will help. Perhaps I'm just hopelessly optimistic?
. Seriously though, just get a sufficiently large target and as far away as you reasonably can. 
