Thom Hogan Declares D800 "Not Recommended"

[ShunCheung]

<p>Rodeo Joe writes:</p>

<blockquote>

<p>As far as I can see my D800 has none of the "widely reported" faults. Infinity focus is bang on with all my MF lenses, AF points act as expected, the mirror is pristine, the battery supplied wasn't on the blacklist and all the firmware is right up to date.</p>

</blockquote>

<p>Rodeo Joe and Andrew Garrard, I causion against celebrating too soon. Since the two of you have brand new D800 cameras, why don't you use them for a few days and look for any AF issues as well as other issues. If after a week there is indeed no problem, please come back and let us all know.</p>

<p>It is not only Thom Hogan; I see a few other internet bloggers are also eager to discuss D800 AF problems. While they try to describe this as a wide-spread problem, I have yet to see a lot of concrete data to back up such claims. It certainly would be nice if someone had the resources to buy 100 D800 from different outlets around the world and check them all out; obviously that is not very realistic.</p>

<p>Checking this week's Nikon Wedneday image thread, roughly 10% of the images (6 out of about 60) were captured with the D800. Clearly quite a few people on this forum have one or two D800. I find it interesting that very few people have posted personal experience about AF issues on their D800. While I am sure <a name="00aeNt"></a><a href="../photodb/user?user_id=2292967">William Scott</a> is not the only person who has experience AF issues, it looks more like the majority of D800 are free of any AF issues, at least among posters to this forum/this thread.</p>

[jeremy_wakefield]

<p>I don't know how typical my experience is. It is possible I had two cameras from a "bad" batch. I received mine on a Friday and tested it over the weekend. When it was found to be faulty I phoned the retailer who put one aside for me that day so both of these came from the same shipment to the dealer.<br>

I know of other cases in the UK of cameras returned to Nikon for repair with mixed results. Some seem to get the issue fixed quickly and without any problem whereas others get the run around. I know this is anecdotal but my own experience led me to wait for a while.<br>

The other worry was the on the second body the centre point AF was pretty erratic and failed to focus correctly - and I mean it was a long, long way out - on several shots. I'd estimate 15% of the time. This was on my 50mm f1.4 lens wide open. The same lens has worked without problem with my D700 and D300 cameras for a while before but I suppose it might be possible it has developed a fault and that this second issue was lens related. (I will keep an eye on it for a while)<br>

<br />All I know for sure is that both bodies I was sent had the AF problem.<br>

I will certainly be looking for a D800 once Nikon sort it out though. If I stuck to the centre to right points or Live View focus, it was absolutely amazing. I was really gutted at having to send them back.<br>

William</p>

[elliot1]

<p>I have had my D800 for a couple of weeks and taken about 1000 images at this point - the AF of the D800 is extremely fast and accurate with every lens I have used (5 different ones at this point). My D3 is not going to see much use anymore.</p>

<p>William, FWIW, I ordered my D800 around the same time you did.</p>

[Andrew Garrard]

<p>Shun: For better or worse, I'll report back as soon as I've had time to test my camera (and do my best to ensure that any problems aren't down to my own incompetence). I've not been home from work early enough to have the chance to turn the thing on yet! (Focus tests in a dimly-incandescent-lit lounge at 2am don't seem worthwhile.)<br />

<br />

Ilkka: I belatedly realised that you're probably right. Teach me not to be snooty about the 50 f/1.8 AF-D. :-) I might bear this in mind when thinking about using mine on the D800, especially in DX mode.<br />

<br />

I'm getting the impression that Nikon UK are a little busy - I've not yet been given a service estimate for the check on the 135 f/2 DC I dropped off last week, but to be fair the only urgency will be if they release a replacement at Photokina before I sell it. I hope my D800E's okay, but it's going to be at least a month before I get the chance to send it in if it's not.</p>

[rodeo_joe1]

<p>Shun, Ilkka, I spent the major part of today checking out a fair proportion of my lenses on the D800. What came out of it was that at maximum zoom in liveview the depth-of-field is close to non-existent, and even with a long-throw Ai Manual focus lens the slightest twitch of my hand was enough to throw the lens off its best focus - but you probably know that already. Add into the mix an f/1.4 lens plus a bit of mechanical backlash and it's easy to see how <em>any</em> AF system might occasionally misfocus.</p>

<p>So, yes, I am seeing some "out-of-focus" shots when relying on AF or focus confirmation, and it just might be slightly worse when placing the focus point to the left of centre. BUT, given the precise job the camera has to do, and the vaguary of wide-aperture lenses I don't think it's too much of an issue. If precise focus is essential then Liveview is the only option, and that went for the D700 as well.</p>

<p>What I'm seeing is that the better the lens, the better the AF performs. Any lens that exhibits a noticeable degree of spherical aberration wide-open throws the AF system off. Then again, the same degree of SA makes it near impossible to accurately focus the lens by eye too. Dear me, AF technology turns out to be no more accurate than the unaided human eye - well boo hoo! However, it is a LOT faster.</p>

[ilkka_nissila]

<p>Andrew, given the relatively low cost of the 50/1.8G AF-S there isn't that much reason to continue using the D version any more (on a D800E or other digital camera). The 1.8G has improved image quality, autofocus and autofocusing over the D version and it's not expensive. I used the 1.8D for many years (from f/2.5 to f/11) and always got good results with it; I never found corner softness to be a problem in my photography. However I was annoyed by the barrel wobble and AF noise and got the 1.4G AF-S and that has been my 50mm of choice since. However, it is more expensive and has quite a bit of barrel distortion. Today I would just get the 1.8G. The 1.8G also has more barrel distortion than the 1.8D (less than 1.4G though) but on the other hand the D800(E) has automatic distortion correction which you can use to check the image in playback to see that everything important is in the frame after distortion correction, if you use it for e.g. architectural subjects. But if your version doesn't wobble and if the AF noise doesn't bother you I don't see any reason not to use it (from about f/2.5 towards smaller apertures) - the lens is small and I always got away with it well. But maybe your subjects and requirements are different from mine. Thankfully today Nikon offers several modern options and prices in this segment haven't gone up as much as they have in some other categories of lenses.</p>

[ShunCheung]

<blockquote>

<p>Shun: For better or worse, I'll report back as soon as I've had time to test my camera (and do my best to ensure that any problems aren't down to my own incompetence). I've not been home from work early enough to have the chance to turn the thing on yet!</p>

</blockquote>

<p>Yesterday, I was reading D800 user reviews on Amazon.com, and one guy is bitterly complaining about the D800 and gives it 1 star on a scale of 5. It turns out that the D800 he received is completely non functional. However, since he didn't have any memory card to test it with, he let the 14-day exchange period pass without testing his new camera. While it is too bad that he got a disfunctional D800, the D800 can take either an SD or a CF memory card, the two most common types; it isn't like you must use an XQD card or something. Not having a card to ckeck our your new camera for 2 weeks is pretty lame.</p>

<p>So Andrew, I sure hope that you won't run into that kind of "problem." IMO that guy mainly has himself to blame. As far as I can tell, most people on this forum have received perfectly working D800 cameras without any AF issue. And just in case you do have a defective one, the D800 is no longer difficult to find so that you should be able to get an exchange from the store quickly; you don't even need to deal with Nikon warranty service. Now that we are all aware of this potential AF problem, it should take no more than an hour or two to evaluate that particular issue on a brand new D800.</p>

<p>However, the D800E may be a little more difficult as it still takes longer to locate a replacement should you run into problems. For the D800, I see not reason why people need to avoid it as long as you are dealing with a store that allows you to exchange a defective camera, within 7 to 14 days.</p>

<p>Having said that, the D800/D800E is certainly not for everybody. Now I have used the D800/D800E extensively, I continue to use my D700 and D7000 whenever I can. 12MP is more than enough in most occasions for me that I would much rather have those smaller RAW files from the D700, and of course in some situations I prefer DX.</p>

[Andrew Garrard]

<p>Ilkka: An f/1.8 G is on my list. :-) Thank you for the recommendation - I'll give it higher priority. I'm not that much of a fan of 50mm, otherwise the price for one of the f/1.4 lenses might be worth the optical issues that I dislike, but - as you say - the new f/1.8 is cheap enough to consider sooner rather than later. Although I may be a while paying off my D800. (I also have a Sigma 150mm, a Zeiss 21mm and a Samyang 35mm on my "to buy" list. Well, and a 400 f/2.8, but that one might have to wait a bit longer.)<br />

<br />

Shun: Thanks for the warning. I'll try to give my D800 a go tonight or tomorrow (I do have cards). I've been bitten before - I got a one-ring 80-200 that I didn't give a serious work out for about three months, and there's something quite seriously wrong with it. I didn't even ask the supplier if I could return it after that long, so I'm treating it as a paperweight until I can sell it for spares (it's not worth the repair cost). The perils of a busy life and procrastination.<br />

<br />

Joe: I gather that, when the left focus point is off, it's <i>very</i> off. Sounds like you have a good sample - glad to hear it!</p>

[elliot1]

<p>That Amazon reviewer said he was getting an ERR code with his memory card, <em>"</em><em>the CF card would only give me a err code".</em> I have discovered that one 16gb compact flash card I have will not format in the D800 even though it has always worked fine in my D3. I have owned it for about 8 months so it is a relatively new card. So if that was the only card I had and I tried it in my D800, my initial reaction would be that the camera is defective.</p>

<p>He also went on to state that Nikon's response was "juvenile such as was I using the right card". It is possible that the card he purchased was not compatible with the D800 and that his camera was/is fine. And that Nikon's response was the correct one. Who knows. The whole story sounds fishy to me.</p>

 

[western_isles]

<p>I looked at Thom's article and he makes a very good point "Most Nikon users are quite loyal. But that loyalty gets strained when Nikon goes silent instead of acknowledging an issue." I had a similar experience with Capture NX2 and the silence from Nikon was deafening until I sent a letter, snail mail, to the Chairman of the Nikon board. Only then did a member of the senior management at Nikon write and apologise for failing to respond to at least two letters. However, they still refuse to acknowledge they have problems with the software! The problem with the D800 appears to be exactly the same scenario. Why do they ignore their loyal customers?<br>

We all know that Nikon make great products, that's why we buy them. However, things sometimes go wrong in all walks of life and Nikon should be big enough to acknowledge that something has gone wrong and fix it. They wont lose customers by doing this. Silence however will drive new customers away which will affect sales, market share, profit...........<br>

Come on Nikon, start treating your customer base fairly. </p>

[rodeo_joe1]

<p>Andrew, I can heartily recommend the Samyang 35mm f/1.4. It's razor sharp from wide-open in the frame centre. The edges and corners do lag behind a bit, mainly due to some field curvature - so you can have sharp centre OR sharp corners at f/1.4, but not both. By f/4 the centre is ridiculously sharp, and the corners acceptable in the same plane. In other words it's not a copying lens, but if you need f/1.4 without a trace of SA in the centre frame, and with good flare resistance and low CAs; this is your lens.</p>

<p>Below is a sample from the frame-centre of the Samyang 35mm @ f/1.4, compared to my old 50mm f/1.4 Ai Nikkor wide-open. Subject is a boring newspaper small ads column I'm afraid. Bear in mind that those clips only show about 0.1% of the D800's image area! (Note to self - must make more use of the camera's built-in virtual horizon and plumb line.)</p><div></div>

[Andrew Garrard]

<p>Thanks, RJ. The field curvature is news to me (I may have missed it in reviews, or they may not have mentioned it) but that's good enough for my needs in a fast moderate wide-angle. Though I have to say the new 28 f/1.8 improves Nikon's line-up here - I'm probably still after the Samyang, but I definitely have a decision to make.<br />

<br />

For what it's worth, the two batteries I have for my D800E aren't on the recall list, and my camera seems fine so far. But since my initial test are of a television from the far side of the room with a 200 f/2 (I can see the sub-pixels), sensor points will have to wait for the morning and daylight. I wanted my first shots to be of something interesting, so the entrance of team GB at the Olympics was it...<br />

<br />

By the way, Thom does have a point about moving cheese - switching the + and - buttons <i>is</i> annoying, especially since I'm expecting still to use my D700. I'd no idea I used muscle memory for them. Oh well, no biggie.</p>

[richard__]

<p>Because someday I might possibly want the D800E Autofocus to work correctly, I thought I should test the new camera now, using Thom's suggestions on his <a href="http://www.bythom.com/D800autofocus.htm">website</a>.<br /><br />I became quite stressed using the <a href="http://www.bythom.com/Images/aftest.jpg">pattern</a> he displays there, when both my right and left sensor completely failed to achieve any acceptable focus in Live View (it likely fails in the center too, but I did not test that.)<br /><br />It does however achieve proper focus when I use my usual target which, if I read his book correctly, unlike his, should be expected to work with the trial and error contrast detection used during Live View.<br>

<br />Note too in using this pattern, it becomes rather easy to demonstrate the problem moire property.<br /><br />So either I misinterpreted his simple testing procedure OR I need to return this camera now if I ever expect to have one where autofocus works reliably in all situations.</p>

<p>That target's image center looks like this and is representative of the pattern I usually use:</p>

<div></div>

[elliot1]

<p>Richard, has your camera been focusing reliably aside from Thom's test?</p>

<p>This is the testing procedure Per Thom's article: <em>"I'd argue that you should be placing your camera at a "<strong>normal</strong>" shooting distance. Some people are testing<strong> far too close</strong>. 6-10 feet (2-3m) sounds about right to me. Your camera needs to be <strong>level and parallel</strong> to the wall. <strong>That step alone can trip a lot of people up</strong>. </em><em><strong>Assuming your setup and test charts are appropriate</strong>, the testing itself is simple enough"</em><br>

<em> </em><br>

<em>"<em><strong>Assuming your setup and test charts are appropriate..." </strong> </em></em>Hmmm... That is quite an assumption! Does anyone wonder why perhaps cameras are failing his test? I am not questioning whether some cameras are not focusing correctly. I am sure some are not. But I also wonder how many are failing this test simply because the setup and test charts are NOT set up appropriately or simple user error of some kind.</p>

<p>In any case, I am certain Thom has generated a tremendous amount of interest in his website over this. Just look how much interest this one thread has generated here!</p>

 

<p > </p>

[Peter_in_PA]

<blockquote>

<p>I became quite stressed using the <a href="http://www.bythom.com/Images/aftest.jpg" rel="nofollow" target="_blank">pattern</a> he displays there, when both my right and left sensor completely failed to achieve any acceptable focus in Live View (it likely fails in the center too, but I did not test that.)</p>

</blockquote>

<p>Doesn't live view do it's "own thing" with AF, not using the non-Live View (conventional) AF sensors?</p>

[Andrew Garrard]

<p>Peter: You're correct, live view just uses contrast detection on the main sensor. At least, it does on the D800 - if it's got the "hand-held mode" (flick the mirror down and use the phase-detect sensors) that the D700 supported, I can't find it.<br />

<br />

For what it's worth, I've now had the chance to do some testing. I only had the chance to use the tripod for some of it, and I'm not going to call it exhaustive, but:<br />

1) My D800 seems to focus almost perfectly with any of its AF sensors using my 200 f/2. This may be expected, because it's a telephoto lens and the problem seems to be with wides. However, it's the only AF lens I own that I could claim is reasonably sharp away from the centre at this aperture.<br />

2) I tried the 50mm f/1.8 AF-D, as I believe someone said Nikon use for calibration. Wide open, the optical aberrations are bad enough that I can't tell whether the focus is nailed; live view has the same problem.<br />

3) I just bought a 50mm f/1.8 AF-S G. It's better, but still not all that tack sharp wide open; you pays your money... However much it may be off perfect focus, it's about the same on both sides. I didn't have my tripod available for doing this.<br />

4) I've just given my 14-24 a go at its wide end - again, I don't have a tripod to hand. I <i>think</i> all is well: I had some trouble locking focus, especially on a blue/white target, and the left sensor seemed to struggle more than the right, but some tests with higher contrast subjects were fine so I don't think it's got a fixed offset.<br />

<br />

So I think my D800E is free from the autofocus problem. At least, if it <i>has</i> the problem, it's not obvious - and my understanding is that when there's a problem, it usually <i>is</i> obvious. Yay for me.<br />

<br />

I now need to go through all my lenses and AF-fine-tune them. And upgrade my copy of Photoshop so I can read the raw files...</p>

[richard__]

<p>Personally I don't think anything is wrong with the camera. But I don't know that as I don't have any big need to actually use the autofocus system, especially so having just received it. I don't have any problem with the camera and the test chart pattern I have used for lens testing in the past. It's only when I use Thom's simple <a href="http://www.bythom.com/Images/aftest.jpg">pattern</a> to test that a complete failure arises in the auto-focus system during Live View mode.</p>

<p>From what I can tell, Thom's simple instructions (which I have followed as best I could*) are not completely clear enough especially when he suggests using that simple pattern for the testing in both situations, which definately invoke differing behavior by the camera's autofocus system in Live View mode. (And additionally, why would I even use that, since I don't see a big need for autofocus normally ?) My reading of the D800 information I have seen is that Thom's pattern suggestion for the live view auto-focus testing is likely incomplete.</p>

<p>So really my comment should suggest my real question is, who else has done his test successfully ?</p>

<p>* My distance was +5', and I created a larger copy of Thom's pattern, using his image sized directly and only copying it about so it would fill most of a letter size sheet which I could print out. That way it was easier to place the focus point(s) onto the pattern he suggests, since his original size is I think too small for the test. I have uploaded it <a href="http://www.henleygraphics.com/CurrVersion/D800aftestSheet.jpg">here</a> if anyone wants to use it. Possibly I need to back off to 10', BUT I'm not sure doing that helps me understand that I have a problem which I don't think exists with this specific camera.</p>

[Andrew Garrard]

<p>Richard: I printed Thom's pattern (smallish, replicated a few times around a sheet of A4). It didn't do me that much good with my 50 f/1.8 AF-D - outdoors in direct sunlight the bloom from the white paper messed with any accurate edges, and inside there was still too much (probably) coma.<br />

<br />

The other tests I performed weren't with his test chart. I agree that he's not clear about how large to print the test pattern or how far away to be, but my feeling is that if I can't focus on a single sharp edge there's probably something wrong anyway, so I was only going to come back to the pattern if I had issues with a simpler test. I just pointed the camera's autofocus sensor at a sharp edge of the right orientation from a reasonable distance away (6' or so with a 50mm, mostly closer with the 14-24). I manually focussed to infinity between shots (less fiddly with AF-S!) and tried several attempts both with phase-detect autofocus and with live view. Live view is sometimes clearly a bit better for me, but there's not much side-to-side difference.<br />

<br />

I've only checked the 50 AF-S and 14-24 on image playback, zoomed in past 100%. I looked at the 200 f/2 and the 50 AF-D JPEG output on a computer.<br />

<br />

I'm not sure that my testing is authoritative, but as far as I can tell my camera's okay - I'm confident enough that, if there's any problem, it's minor and not the systematic problem that seems to have been reported. I've submitted my results to Thom's poll (as "acceptable").<br />

<br />

Why would one use live view autofocus? Because it's normally more accurate and controllable, if there's time to use it. Why would <i>you</i> use it? That depends why you're not using autofocus already - it may save you time. Unless something is horribly broken, manual focus at the same location would suffice - I suspect the instructions are just a quick way to ensure that the same bit of the image is used for focus. Of course we expect differing behaviour - the point of the test is that the problem is with the phase-detect autofocus sensors, not live view. I'm surprised that your live view autofocus failed to cope: it only has to move the lens around until it finds the sharpest edges, and that should be relatively bullet-proof. It certainly worked for me.<br />

<br />

Summary: I "sort of" did his test, but I think the special pattern is overkill.</p>

[ShunCheung]

<blockquote>

<p>It's only when I use Thom's simple <a href="http://www.bythom.com/Images/aftest.jpg" rel="nofollow" target="_blank">pattern</a> to test that a complete failure arises in the auto-focus system during <strong>Live View</strong> mode.</p>

</blockquote>

<p>Richard Henley, the entire discussion about left AF point problems with the D800 is only about using phase-detect AF when you are using the traditional viewfinder. That AF system is at the bottom of the camera and requires a secondary mirror behind the main mirror to bounce light down to the AF module.</p>

<p>When you use live view, the mirror is up and that phase-detect AF system is not used. Therefore, what you are doing in live view mode has absolutely nothing to do with what Thom Hogan is discussing (concerning the left AF point problem). In the live view mode, there is no 51 AF points any more; you can AF at any point on the sensor and it uses contrast-detect AF, which is something completely different.</p>

<p>Peter Hamm has already pointed that out earlier.</p>

<P>

See the image below showing that the center of the mirror is semi transparant and there is a secondary mirror behind it to redirect light down to the AF module at the bottom of the camera.

<BR>

<BR>

<CENTER>

<IMG SRC="http://static.photo.net/attachments/bboard/00a/00aX3c-476151584.jpg">

</CENTER>

</P>

[richard__]

<p>Well it is part of Thom's suggested testing, why it is I don't know or why it is relevant:</p>

<p>>> <em>Set the camera to Live View. ...</em><br>

<em>>> They should all seem to be in relatively good focus. If they're not, do not pass go, do not collect your $200.<br /></em><br>

Otherwise as I hoped I suggested, it doesn't appear to be one that is very useful to discerning the actual concerns owners have. And by lacking in details, but with enough said to worry anyone right off if it fails there.</p>

<p>Because I neglected to mention that I don't have any problem with phase detection autofocus and his pattern either, I was definitely not clear enough on that point.</p>

[Andrew Garrard]

<p>Richard: I think he's just trying to provide a good target on which the autofocus sensor should lock. If I'd had trouble with a convenient sharp edge, I might have tried harder with a pattern like this.<br />

<br />

As for why you should try it with the contrast-detect autofocus, I suspect that's a way of checking that the combination of your test pattern, lighting conditions and your lens's performance isn't such that the autofocus can't be expected to see what it's trying to focus on (or that you accidentally left the lens in manual focus mode). Since we're trying to tell whether the autofocus is out of alignment, rather than whether it works, it's just confirmation that any failures aren't down to the testing conditions - there seem to have been a lot of false positives with paranoid consumers anyway. If you know what you're doing, don't sweat it - I'm sure Thom's just compensating for people who don't know what they're looking for.</p>

[jeremy_wakefield]

<p>You use live view to get the best possible image given any shortcomings of your lens etc. This should be accurate and the standard to which you can reasonably compare your "viewfinder" AF test results. In other words the Live View image isn't the test, it simply becomes your reference point against which you actually test the AF points</p>

<p>William</p>

[Kent Shafer]

<p>Shoot! I was perfectly happy with my D800E until until Tom Hogan and all you other spoilsports convinced me I should test it, and I now find it does seem to have the dreaded left focus problem. "Oh, bother," as Winnie the Pooh said.</p>

<p>Here's a 100% crop comparing two images taken of a test target with the 50mm f/1.4 AF-D lens at f/1.4. The one on the left was focused using the left center focusing spot in the viewfinder, and the one on the right was focused on the same area (the horizontal bar pattern provided by Thom Hogan) using the contrast dection autofocus in Live View. There's a pretty clear discrepancy here, but it's nothing compared to the 14mm test I'll show in the following post.</p><div></div>

[Kent Shafer]

<p>And here are comparisons using the same procedure with the 14-24mm zoom at f/2.8. Yikes!</p><div></div>

[Kent Shafer]

<p>I did the same procedure with the same lenses using the center spot and the right-center spot, and everything matched perfectly so far as I can tell. The problem is just on the left.</p>

<p>The local shop where I bought the camera says I can either bring it in and have them send it to Nikon or do it myself. If they do it, they'll charge me a US$25 shipping fee, which they say Nikon has recently started charging them for warranty repairs. That's probably what I will do since it's still cheaper than shipping, insurance, and professional packaging at the UPS Store.</p>