snapsy
Offline
• • • • • •
Upload & Sell: On
|
 p.3 #8 · p.3 #8 · Dot-Tune: Autofocus Fine Tuning in under 5 minutes | |
Todd Moon wrote:
This method is really nice (most of the time). Thanks for posting it! I had previously used a value of -10 for my 70-200mm f/4 IS, but this method resulted in -5 to be the best setting (at 200mm).
You don't have to check the viewfinder for the green dot if you have focus beep enabled. That's much easier than looking through the viewfinder. Just make sure you press the shutter a few times to make sure it beeps consistently.
I said "most of the time" up above because I found a situation where this technique fails. I put my 28-135mm lens on my camera and on the positive end it beeped at -7, but failed to beep at -6, so I marked down -7. Then I went to go test the far negative end and it beeped even at -20 which is the MFA limit of my 50D. (I'm not sure if other cameras go higher.) It's very unlikely that -20 is the actual limit that this lens would give me focus confirmation. It's more likely that I simply hit the limit of the MFA setting before learning the actual point where I lose focus confirmation.
To test this I set the MFA to -13 (half way between -20 and -7), turned on AF, manually focused to infinity, half pressed the shutter release, and then checked the focus in 10x Live View. It looked good. I did this about 6 more times and it would occasionally back-focus. So I did the standard "trial and error" MFA discovery technique and found that every shot looked good at -16. Even that may not be the perfect setting, but that's the first value I tried, and in my test of about 8 more focus attempts, every one looked very crisp in 10x LV.
Now, on my 70-200, the range of MFA values that gave focus confirmation was -13 to +3. That's a total range of 16. If we assume that every lens has the same width of values, then the my 28-135's range should have been -23 to -7. The middle of that being -15, which is pretty darn close to the -16 that I kept. So, for problem lenses, perhaps if you simply find one end of the range, you can just add or subtract 8 to it you have your value. However, I don't really know if that range of 16 is specific to that lens, or my camera, or not. More testing with more lenses is needed.
In conclusion, this technique is super fast and easy as long as your lens's optimal MFA value is not less than 8 away from one extreme. If it is, you have more work to do.
EDIT: The acceptable MFA range on my 50mm f/1.4 was -1 to +15. Again a range of 16. I previously had this lens set at +10, but now I set it to +7. I love how fast and easy this method is!
EDIT2: The acceptable MFA range of my 17-55mm f/2.8 at 55mm was -3 to +13. I'm definitely seeing a pattern here!  At least on my 50D, the range of acceptable values is always 16 wide.
Every test was done at 50x the focal length distance from my target....Show more →
Thanks for the feedback! Regarding your 28-135mm, I agree, you likely received a false confirmation @ +7. One aspect of this technique which is hard to convey in words is how you have to develop a feel for what represents a consistent and reliable confirmation vs one that is marginal. The feel for this comes after using the procedure a few times. On both Canon and Nikon bodies you have to half-press the shutter (or AF-ON) several times and sometimes hold it for up to 5 seconds to see if the confirmation waivers. Additionally, on Canon bodies I've found that if there's any delay between half-pressing and the confirmation then it's likely not a confirmed AF value, even when the delay is really small. On Nikon bodies, which have a front/back rangefinder, you get more nuanced feedback about a marginal AF value from a quick flicker of one of the rangefinder arrows. With both brands though there's definitely some finesse involved.
For cases where one end of the tune range still shows consistent confirmation even at +20 or -20, you can slightly adjust the focus of the lens to shift the range closer to the center, even if that means that adjustment puts the subject OOF. With the range centered you can then determine what the range is...then apply that range to the AF tune values you originally got with the subject in perfect focus. For example, if the range is +5 to >+20, shift the focus a bit until you get a range that fits...like to -15 to +15, which will let you determine the width of the range (30 in this case), then you can apply that range to the original +5 to >+20.
You likely can't rely on the width of the range being the same across lenses on a given body. That width is determined by several factors, including DOF and target contrast. On my D800, the Sigma 35mm f/1.4 has a range of only 2 AF tune points!, whereas my 50G f/1.4 has a range of 8 points.
|
