01-09-2019 05:18 PM
If I understand correctly you are getting your best result at 10 units of AFMA (worse results on either side of 10) but still not sharp at 10. That’s strange behavior.
01-09-2019 06:24 PM
Try this, using single AF point focus on a bright contrasty subject:
1. Focus with Live View. Note position of focus ring/scale.
2. Switch off Live View and half-press shutter button to focus.
3. See which way the lens focus ring moves - towards or away in distance from the LV position.
4. Move AFMA adjustment in opposite direction. If focus moved in direction of infinity add negative AFMA or vice versa
5. Repeat until lens doesn’t move between LV and viewfinder focus.
01-09-2019 08:00 PM - edited 01-09-2019 08:01 PM
Liveview focus and viewfinder focus use completely different focus systems. On a 5D IV it's a little different than older cameras because the 5D IV has Dual-Pixel CMOS AF.
In any case, Liveview focuses the image directly ON the sensor itself. The reflex mirror is "up" so the light can hit the sensor. When you shoot via the viewfinder, the reflex mirror is down ... light can't reach the sensor. The mirror bounces most light up to the viewfinder screen ... but there's a semi-silvered area in the middle of the mirror that lets some light pass through. A second (smaller) mirror is hiding behind the first mirror and this mirror is angled to bounce light down to the floor of the camera ... this is where the phase-detect AF sensors are located.
The theory is that if the distance light has to travel to go through the first mirror, hit the sencond mirror, bounce down, and reach the AF sensors is *precisely* the same distance that the same light would travel if the mirror were out of the way so the light could go straight to the sensor, then a focused image on the AF sensors would mean a focused shot on the camera imaging sensor.
But the sensor has to be shimmed into position to locate it at *precisely* the right point. And it turns out different lenses can also have different effects.
SO... the 5D IV (and most high end bodies) have Auto-Focus Micro-Adjustment that lets you tweak the focus to tell the camera to focus slightly closer or slightly farther than it otherwise would.
There are focus calibration tools (X-Rite makes one called the "LensCal"). LensAlign is another popular tool. They both work very similar to each other.
The idea is there is a flat target that you use for focus.
Adjacent to the flat target is something resembling a yardstick or a ruler .. except it is resting at an angle (like a ramp). It has a scale on it.
You point the camera at the FLAT target (not the ramp) and take the shot. But when you pull the photo into your computer, you inspect the focus on the ramp ... to find the spot where you think focus is the best. This might be either closer or farther than the flat target (or it might be bang-on accurate).
As you see how these things work... you could acctually just use a flat target and rest an actual yardstick at a ramped angle immediately adjacent to your target (make sure the AF point is actually ON the target ... not the yardstick.)
You should be able to adjust the camera for each lens so that it nails focus. I actually suggest deliberately de-focusing the lens so that focus is deliberately closer than the target... make the camera adjust focus outward to achieve focus and take the shot. Repeat this a dozen times. THEN... deliberatetly run the focus out too far (toward infinity) and make it adjust focus inward to achieve focus and take the shot. Also repeat a dozen or so times.
Just a single sample could just have a freak incident... but if you take a dozen samples in each direction you *should* be able to see a trend and you want to adjust the camera based on the sweet spot in that trend.
If focus is all over the place (there is no trend) ... try a different lens (the lens may have sloppy mechanics and require service).
If the focus is all over the place regardless of lens... send in the camera (and you may want to send the lenses along).
If the camera is so far out of focus that even the limits of the AFMA can't correct for it ... send the camera in for service (sensor probably needs to be re-shimmed to the correct spot.)
01-09-2019 08:28 PM
yes best result was at plus 4 but not as good as in live view
Maybe you do you need an AFMA adjustment, but I would doubt it with new gear. But, who knows?
Follow Tim’s suggestions on setting up a target. You need to be about 25-50x the focal length, so you seem to be at a good distance. I think 50x would be better, though. For super telephoto lenses, I like to use either a yardstick or a tape measure.
Check the focus range settings on the lens. You can get soft images if it is set to the wrong range. Which way should you adjust the lens, with a positive or negative value? The answer is pretty simple. Think of the AFMA value as a correction to the focusing distance.
If the lens is back focusing, then enter a negative value to bring the distance to the plane of focus closer. If the lens is front focusing, then enter a positive value to increase the distance from the camera to the plane of focus.
I have found that the “Dot Tune” video on YouTube to be very simple to use. It takes all of the guess work out of making an AFMA. Just take numerous sample shots, and calculate an average. Hope this helps.
01-10-2019 02:01 AM
Thanks guys, I will have to try this at the weekend as it is dark when I go to work and dark when I come back.
The camera is new but the lens is a few years old
thanks again, I will get back
01-10-2019 07:19 AM