Acuracy of AF focus with fast lenses

[jon_savage]

<p>Hi All,<br>

I was after some advice to help set my expectations. I have a D90 (13 months old) and the 35mm 1.8 (since last Sunday) and the 50mm 1.8 (couple of months).<br>

All was fine until I bought the 50mm. I then became suspicious that it was short focusing. Single centre point focus on "AF-S" AF mode. I took the camera and lens together back the the shop (I'm in the UK) and they sent them away to Nikon to be looked at under warranty. They came back 4 weeks later with no fault found. So I thought I was just being fussy. Just fast lens focus syndrome? - related to pixel peeping noise/sharpness syndrome I guessed.<br>

Anyway Chrismas school play so out with the 50mm lens. All but one photo of him soft, all the sharp areas appeared to be ~1ft plus in front of him. So back to the shop. By this time I was after a the 35mm so I wanted to try this before I bought it and at the same time discuss the focus. This time the boss was in and with no hessiation after trying the camera and lens offered to swap the 50mm for another when they get them back in stock. The 35mm looked to be focussing better than the 50mm so I bought it.<br>

Today I went in to exchange the 50mm so before I did I tried it. I was not convinced it was any better than the one I had so I didn't exchange. I tried my lenses on another new D90 and that wasn't a whole lot better. My test piece is a bit of flat lego but it's easy to demonstarte on any flat sheet with text on it.<br>

So at this point I just thought D90's and the 50mm's shallowest DOF are just hard to manage. Then I tried the same test using live view to focus. With both the 35 and 50 the focus point was spot on and way better than any of the new/old D90/lens combinations I tried.<br>

So my question is is this typical and expected focusing behaviour and are my expectations too high? I understand I could have both lens and camera within tolerance but both at extreems and the sum is noticable.<br>

I assume live view uses feedback so will always be good. Does the standard phase shift method just measure how far out of focus it is and move the lens a predicted amount (and not check again) - so the lens response could need calibrating? Then there's the alignment of the mirrors and focus points - they could be out?<br>

Is there a recommended way to get everything tuned back to nominal to improve (other than uppgarde to a D300 with its adjustment feature!)? Should I expect that under warranty even if they are both individually "in spec"? Ultimately it's a system I need to be right.<br>

Sorry for the rambling but any advice would be appreciated.<br>

Regards, Jon<br>

Example photos to follow..</p>

<p> </p><div></div>

[jon_savage]

<p>100% crops of centre of images. The focus point is intended to be the centre red circle with the red part on its right. LVAF = live view method, the only way I can get exactly what I want in focus.</p>

[jon_savage]

<p>Sorry, labelled my files the wrong way round, this should be the 50mm one...</p><div></div>

[stefan_g]

<p>Hi Jon, I would expect Nikon to get the adjustment right. At least that's what they did for my D70 a few years back ( http://www.photo.net/digital-camera-forum/00EszD ). When I sent my camera in I told them what I thought the magnitude of the misalignment was. But even without that they should be able to figure it out.<br>

Good luck, Stefan</p>

[raymond bradlau]

<p>My educated guess is that your red AF marker is larger than the actual point you are trying to focus on.</p>

<p> </p>

[jim_gardner4]

<p>Jon, there was a very interesting article posted on P.net a week or so ago about what you have mentioned and matching bodies with lenses.<br>

A friend of mine is a proffesional photographer in London and I mentioned short focusing to him a few years ago. He thought it was quite ammusing that I was using AF ( I should point out this was a few years ago and things may have moved on since then) and if I was going to shoot at 2.8 / 200mm I must focus manually.<br>

AF is good but not as good as the human eye.<br>

I cant remember who posted the article I mentioned but if you can find it somehow it is well worth reading.</p>

[jon_savage]

<p>Thanks for the responses. I've attached full view pictures with the size of the red AF squares shown so it may give a better idea of the issue.</p>

<p>I suspect the 50MM is a bit more front focus biased than the 35mm and the camera also a bit front biased so they add up to what we see.</p>

<p>I guess it boils down to one of the options below:</p>

<p>1) it's as good as it gets & don't worry about it</p>

<p>2) it's typical of manufacturing tolerances and therefore not covered by warranty (that's why the D300 and above get focus tuning functions)</p>

<p>3) it could be better but I'd have to pay to get them tuned/calibrated (and also why the D300 and above get focus tuning functions)</p>

<p>4) it could and should be better and covered by warranty</p>

<p>Anyway it's Christmas so I'd better be off and capturing memories. I just need to remember to lean forward a bit after I've focused!<br>

Merry Christmas all..<br>

Jon</p>

[jon_savage]

<p>Pictures...<br>

35mm focus square<br>

35mm live view square<br>

50mm focus square<br>

50mm live view square</p><div></div>

[jon_savage]

<p>another</p><div></div>

[jon_savage]

<p>50mm...</p><div></div>

[jon_savage]

<p>last one...</p><div></div>

[raymond bradlau]

<p>I vote option 5, your AF point is larger then the part of the image you want to focus on..... so either have Nikon install a mind reading chip or come up with a plan to overcome this when shooting in this type os situation</p>

[jon_savage]

<p>Raymond, Can I check I understand you right? I agree I’m assuming the AF point (the area within the red square in the viewfinder, or the Live View square area) is larger than the point I want to focus on. Why would that be the problem? <br /> <br /> The technique I used was to expect to get something within the AF point in focus as that was where I wanted the plane of focus to be. I’m telling the camera to just make sure something within this one chosen AF square is in focus.<br /> <br /> As long as I have some variation and contrast within the square area there should be something for the AF to work with. i.e. don’t fill the area full of a relatively smooth tone like a painted wall or a dark shadowy area. <br /> <br /> The AF doesn’t struggle and hunt to try to lock on, it’s straight to a position it likes. The Live View results are what I would expect, something within the area in focus.<br /> <br /> But what ever the reason I think your saying it’s as good as it gets and to plan around it. For photographing that all important lego close up (!) I can use live view or MF with live view zoomed in. <br /> <br /> Thanks, Jon</p>

[raymond bradlau]

<p>How do you know what you are focusing on if the red focus brackets are considerably larger than your target? Its locking on fast and accurately but to what?</p>

<p>I sometimes get the same focus issues when shooting full length people shots at a very wide aperture obviously I want to get the eyes in focus but my AF point is the size of half there head..... I wish there was a cross hair focus point option but I dont think AF works like that so we used get used to it in these cases </p>