Calibration - is this normal?

dubcat

dubcat

Suspended / Banned
Messages
1,310
Name
Amir
Edit My Images
Yes
Hi, I've never been happy with my Sigma 50mm 1.4 lens. I shoot with it on my Canon 50d. At 1.4 I always get very soft (not pixel peeping soft but genuinely soft) results with a ghost effect.

Today I signed up to the FoCal beta on Mac and calibrated my lens. When I took some test shots along a ruler the camera always seems to be focussing in closer than where the focus point is set by about 2cm. I did many many tests and found this to be the case consistently. I was really frustrated.

I was about to give up. I was sat at my desk. As a last ditch effort I started shooting at the middle of the target from the focal test sheet which was lying on my desk. I pulled them in to lightroom and low and behold most of them seem to be bang on. What is going on? Anyone have any idea?
 
Last edited:
Not exactly sure of what you are doing but if the 'ruler' has a chart beside it then it's the chart you should be focusing on, not the ruler ... if that applies and makes sense :)
 
Gramps, it wasn't. In fact I rested a black ball point pen with a shiney tip next to the point I was focussing at.
 
If I'm understanding you correctly this has been covered many times. If you are focusing directly onto an angled ruler, and with the actual AF point being larger than the indicated focus square, the camera has a choice of places upon which to focus because the AF point covers a range of distances with an angled target. Most usually it will probably focus towards the front in this scenario, but I'm not sure that it will always be the case. The fact is you can't trust tests performed in this fashion.

This is why advice, over and over again (including discussions between you and me in the past), is only to focus onto a target which is parallel to the sensor. You can use an angled rule beside the target from which to better judge whereabouts focus has been achieved, but you do not focus on the ruler itself unless you enjoy wasting your time.

If you perform a Google image search for "lensalign" you will see the geometry of a well designed test setup....

https://www.google.co.uk/search?q=l...kIGoDQ&sqi=2&ved=0CAcQ_AUoAA&biw=1280&bih=685
 
Last edited:
Have a look here for example:-

error in German!
 
Last edited:
Try here:-


[YOUTUBE]t-JrUQO_Z5U[/YOUTUBE]
 
I was not clear - i was focussing on the shiny metal tip of the ball point pen.
The ruler was under it - yellow ruler with black lines, i figured the focus would lock on to the shiney metal tip in preference.
 
I used the FoCal target stuck on a wall. I figured this would be good enough for now. I'll start using the camera and see what happens in real world shooting.
 
I wouldn't count on the camera favouring the shiny tip over some well defined markings on the ruler. Also, if the tip is above the ruler then, depending upon how much, it is going to be closer to the camera than the part of the ruler directly underneath it. Results will at best be uncertain. Why not remove the uncertainty completely?

Focal is good, but that also has requirements for proper setup - the right sort of light, sufficient light, and a solid tripod on a solid floor - not carpet etc.. Even floorboards are susceptible to movement which can throw results off if people are moving around. Depending on focal length you might also need to extend your MLU delay. I use 2 seconds because there is enough vibration with a long lens (say 300mm+) and big magnification to make that necessary. With shorter lenses the default should be enough.
 
When I had a Sigma 50 1.4 it had a curious feature where you would calibrate it and get it focussing bang on and as soon as the subject distance changed, the calibration would be out. It could be calibrated perfectly for a given distance, but the lens would front or back focus if you moved further or closer to the target.

Reading online, some people have no problem at all with the lens, some people have all sorts of weird AF issues with it. I rather suspect it plays better with some camera bodies than others. I tried two copies on a 5D2. Its a shame I couldn't get on with it because its a great lens.
 
My Sony 50/1.4 requires slightly different focus calibration depending on the subject/target distance. My solution was to test it at two distances and average out the amount of adjustment to arrive at the best overall setting. It's still slightly out (bf) at 1.4 at mfd, but once you know that its fairly easy to compensate for (in DMF mode) manually.

Oddly, none of my other primes suffer this problem. At least, not noticeably.
 
I will check this distance thing out. I actually read something similar reported from users of the Sigma 35mm f/1.4 on another forum.

I took some test shots post calibration. I sat in the conservatory (loads of light), my daughter decided to pose for the camera and sat still, I upped the ISO to 200 anyway and fired away being really careful to aim the active (centre) focus point at the her eye (specifically trying for where the white of the eye and the iris meet).

I had a few hit's and a lot of misses. I have no idea if this hit rate is normal or not. I have started a thread on that here:

http://www.talkphotography.co.uk/forums/showthread.php?p=5294616#post5294616

I would appreciate your feedback in that thread. Thanks for the help so far.
 
You must log in or register to reply here.
Back
Top