When the camera says move x amount in this direction, it assumes that x is the perfect amount, and the lens will correctly move to this point. It’s fast, but it relies on your measurement being accurate and the lens doing exactly what you tell it to. So one quick measurement can give you everything you need to know – then you drive the lens the right amount in the right direction and voila, perfect focus… Phase-detect AF (PDAF) looks at two paths of light coming to the sensor and with a single measurement can determine both the magnitude (how much you need to adjust the focus) and the direction (towards the camera or towards infinity). The down sides are that CDAF is quite slow and requires good light. The end focus position is probably about as good as you’re going to get (although you can never be quite sure as you have to adjust the lens to check, which might then mess up the focus!). This is a closed-loop, iterative process – closed loop means you adjust and then check your result, and iterative means you repeat the process until you’re happy with the outcome. You repeat this process until you get a sharp image. CDAF works in the same way as a human focusing a lens – you look at how sharp the image is, and if you’re not satisfied with the sharpness you move the focus point a bit in a random direction and see if it gets sharper or not. I'll do some testing with the 85/1.2 which I do have here over the next few days as being very wide that should show some degree of the same issue.If the camera uses purely contrast-detect AF (CDAF), you wouldn’t need any sort of calibration. I would try an EV of around 6 and see if this makes the results more consistent for you. I know the recommendation is to run at EV 10, but in this instance it might actually be causing some issues. I can't do any testing here as I don't have the lens available at the moment, but what I would suggest is that you turn the lighting down (your test was at EV 10). The lens appears to show a high degree of CA (even in the centre) at f/1.2 when there is a slight misfocus, and this has quite an effect on the analysis calculations. They're really thorough and I think I can see a problem. it does seem to be quite a picky lens for FoCal to work - thanks very much for the logs you sent me. The analysis method has been tweaked to try and reduce the effect of any lens image shift and changes in light level, as well as completely reworked "insides" which should make things generally more consistent and reliable.Īs for the specific issue with the 50/1.2. Things are completely mad at the moment with development of FoCal, and there's load of improvements to come in FoCal 1.5. Hi all - sorry for taking some time to get into this thread! Ran it again a few times and got 2 or 3 results at +8 or +9 at 200mm. Sometimes it CLEARLY shows 0 is the best value, when clearly in real shooting it is not. I just wish results were more comparable when tested in slightly difference conditions and distances. Results get better closer to MFD, but at 25-50x FL they can be more "iffy".īut after a day of tweaking, I think I may have got the values needed.
And its not just my copy though either, this is the 2nd 70-200 I have used with the same type of results (on 5D2 and 5D3). Like I said, my 24-105 and 135L calibrate with ease with clear data. Compared with focus consistency as above, and think I got it narrowed down to the right value. Ensuring the resolution recorded was highest on that value. Confirming the best values with focus consistency testing at those values, compared to ones close by. I ended up going with 4W-4T without the TC with my 70-200. Quite a big jump, not sure why it changed so much. 135L switched from +4 from a previous testing to -4 today.
Results look good though with all other lenses.
This is the second time running all my other lenses and all the values changed since the first time I ran them. I get results from 0W-9T to 6W-4T the next. But my 70-200 II is my problem child, with and without the TC. Lighting was +13EV in bright daylight, camera perfectly level and squared up best I could. All between 25x and 50x the focal length. I spent all morning today doing all of my lenses, running them sometimes two or three times at various distances. I like Focal, but get mixed results at times.
0 kommentar(er)
