What level of consistency is the Nikon focusing system going to give in the D800? It seemed so consistent with the D700 I tried at CES/PMA but now massive amounts of doubt are being thrown in...
I would really, really...REALLY...hate...HATE...switching to Nikon for the D800 and finding out that the camera doesn't even focus consistently accurately..
People messing about with "tools" like the lensalign, shooting it from a distance 3-10x shorter than what they actually use and in the wrong light make up for more than 50% of the "my AF is crap" reports.
And then they seem to totally miss out on the fact that the MA feature in the camera is a BODY related adjustment, it has nothing to do with the lenses. The lenses do within themselves contain tables of correction values that the body should apply to the result it finds (after it has done its' own corrections). So if all of your lenses make better with say -10 (tested in VALID circumstances) then the body is probably out by that amount. If there's any difference between lenses, the problem is in the lenses - not in the camera.
If your lenses all need different values, then they will most probably not be consistent over distance either. If you MA for one distance, then they will focus correctly on that distance. No guarantee for shorter or longer distances.
Knowing the communication protocol I also know that writing a CDAF/PDAF correlation would be pretty simple, as long as you restrict the correction to the lens serial no, not the lens model. The problem is that you would then have idiots calibrating their cameras in incandescent light at 50cm shooting distance out shooting in deep shadow 9000K light at 20m distances complaining that the system doesn't work. The manufacturers don't want that (and it's a 100% certain situation) - hence they sit very tight on the calibration software.
theSuede wrote:
People messing about with "tools" like the lensalign, shooting it from a distance 3-10x shorter than what they actually use and in the wrong light make up for more than 50% of the "my AF is crap" reports.
And then they seem to totally miss out on the fact that the MA feature in the camera is a BODY related adjustment, it has nothing to do with the lenses. The lenses do within themselves contain tables of correction values that the body should apply to the result it finds (after it has done its' own corrections). So if all of your lenses make better with say -10 (tested in VALID circumstances) then the body is probably out by that amount. If there's any difference between lenses, the problem is in the lenses - not in the camera.
If your lenses all need different values, then they will most probably not be consistent over distance either. If you MA for one distance, then they will focus correctly on that distance. No guarantee for shorter or longer distances.
Knowing the communication protocol I also know that writing a CDAF/PDAF correlation would be pretty simple, as long as you restrict the correction to the lens serial no, not the lens model. The problem is that you would then have idiots calibrating their cameras in incandescent light at 50cm shooting distance out shooting in deep shadow 9000K light at 20m distances complaining that the system doesn't work. The manufacturers don't want that (and it's a 100% certain situation) - hence they sit very tight on the calibration software....Show more →
Since I was the one who mentioned I own the lensalign I'm going to immediately assume your first paragraph is directed at me... Funny that you have no clue what led me to even break it out and do a quickie test to see if the lens was off, or if I was off... (Oh, it was the lens)
36 MP's is going to do nothing for the majority of photographers buying the camera except show how bad their technique it...
Since this pertains to a recent post I made about my Canon equipment not liking tungsten light...can someone please test their D800 and see if there is any focus shift with any of their lenses under tungsten, fluorescent or other lighting?
jhinkey wrote:
One would think that with the demands of the D800's high MP sensor that the PDAF system would both be more capable of higher accuracy AND be better calibrated right from the factory.
This is one the of the outstanding features of my D800. The AF precision is significantly better than on my 3 series bodies.
form wrote:
Since this pertains to a recent post I made about my Canon equipment not liking tungsten light...can someone please test their D800 and see if there is any focus shift with any of their lenses under tungsten, fluorescent or other lighting?
That was one of the first things I tested, since it's a known problem with the D7000. If the problem remains, it's small enough not to be noticed in any normal situation.
ACNYPhoto wrote:
Since I was the one who mentioned I own the lensalign I'm going to immediately assume your first paragraph is directed at me... Funny that you have no clue what led me to even break it out and do a quickie test to see if the lens was off, or if I was off... (Oh, it was the lens)
36 MP's is going to do nothing for the majority of photographers buying the camera except show how bad their technique it...
The lensalign is used by most users at MFD and indoors in either incandescent or halogen/incandescent light. Properly used (in 5000K+ light if you're going to use the lens outdoors, at low magnification - at least 30x FL, preferably 50x) it's not bad. But the target is a bit to small then...
You mis-understood. The higher MP of the D800 will show AF errors more readily than previous bodies. Therefore it's more important than ever to have a more accurate and better calibrated PDAF system (AF sensors, sub-mirror, etc.) than with previous 12MP or 24MP FX sensors.
I disagree with this, somewhat.
On a per-pixel level, sure the D800 will show errors because the pixels are smaller, therefore any shift in focus or movement, etc. will be more readily seen on a pixel level.
Most of us, though, don't view images on a pixel level - we view them on a whole-image basis. In these instances, the high resolution is going to negate the errors inherent in the shot (assuming they're not atrocious errors). I've performed this "trick" many times before with slight misfocuses with my A850 and now my A77 - downsample the file and post to web and you cannot find the errors.
Now, this assumes a few things - The high-res camera is used like a lower-res version (so that if there are errors in the image, the errors are the 'same' between bodies, even if their representation isn't...aka you didn't use perfect technique on one and not the other), and we're assuming print / view sizes are the same.
Once you get into cropping files heavily, of course things change.
----
In summary, you're only going to be penalized for sloppy technique if you're looking at 100% screens, cropping heavily, or printing much larger than you normally do with a lower-MP camera (and then again, normal non-peeping viewing distances come into play on this last instance, again negating some of the errors, and this doesn't take into fact a lower-resolution camera up-rezzed might not look as good).
