[Marc Sabatella]

There's a difference between saying "some lenses are slightly off while still
within spec" and "it's not reasonable to expect an in focus picture at f/2.8". I'm
simply saying the former.

I agree. And that's actually my point. When potential customers hear about BF/FF problems, they might be imagining cameras and lenses that just cannot get the subject in focus at f/2.8 - because Mike used words almost exactly to that effect. So I'm just trying to clarify that that this is *not* the usual case - the extent to which the system might be "slightly off" should not prevent one from getting in focus pictures at f/2.8.

I think that in some contexts, they still could be used as synonyms, but in
this case a distinct 'hack' existed for the K10D, so it maybe wasn't the
best term to use to describe a nice pretty GUI in the K20D/K7/D300.... etc.

You're right. I'll try to remember to call it kludge from now on :-). Although I'd say the feature itself isn't a kldueg - it serves a very useful purpose in fine-tuning a correctly functioning but not 100% precise system. It's using this to correct a lens that can't focus at f/2.8 I'm calling a kludge.

But if you want to label it a kludge, that's fine, I just don't agree that it's a
kludge any more than, say, monitor calibration is a kludge, or that seat belts
shouldn't be used because they represent a capitulation to all the bad drivers on the road.

The difference here is that with respect to AF, if your camera is so far out of alignment that it can't focus at f/2.8, there is a *proper* (non-kludgy) solution: have the AF sensors adjusted. There is no similar "proper* solution to the problem of poor monitor colors or bad drivers. Although I do see your point with respect to monitor calibration - *in theory*, this is something that should be adjusted in hardware. In practice, it just doesn't work that way, though.

[madbrain]

1200 shots off AA's?

what are these mythical beings?

Eneloops. And I can confirm that I have gotten 1200 shots off one set, in a single day in June, outdoor, without flash, with my K200D.

[schmik]

I agree. And that's actually my point. When potential customers hear about BF/FF problems, they might be imagining cameras and lenses that just cannot get the subject in focus at f/2.8 - because Mike used words almost exactly to that effect. So I'm just trying to clarify that that this is *not* the usual case - the extent to which the system might be "slightly off" should not prevent one from getting in focus pictures at f/2.8.

Maybe... maybe not. I didn't think my words were that harsh but I am jaded by my experiences with AF, DA40LTD (two copies) and pentax bodies. Do canikon do any differently to pentax? from what i hear, "No", they all have their problems.


It's always going to be an issue for someone with a body with no AF adjust. If you have a lense with a tiny bit of backfocus and another with a little front focus? If you send it in to pentax do you think they will calibrate it all for you for free or tell you to deal with it? how long will you be without your gear for? and let's say they do actually calibrate the lenses to be perfect on that body, what happens when you upgrade bodies? I'd say the service guy will use the debug menu to get as good as he can and not open up the lenses for any calibration. Much quicker, easier and cheaper to use the software fix. This will leave you with a compromise.

Most people will never test their gear to this level anyway and don't really need it to be 100% accurate. I'm probably talking about pro spec quality here. The KX is hardly pro-spec. Most people will use the kit lense anyway on the KX and not ever see the issue.

I'm not saying that a KX with a LTD lense won't be able to focus at f/2.8 but I am saying that in my experience that combo may need tweaking to be 100% and that the chances of you getting a 100% accurate combo out of the box are not gauranteed...

mike

[Aristophanes]

Eneloops. And I can confirm that I have gotten 1200 shots off one set, in a single day in June, outdoor, without flash, with my K200D.

I hit 1312 off one set myself over a week worth of shooting. I could have done more but the AF started to act up.

[Marc Sabatella]

Maybe... maybe not. I didn't think my words were that harsh but I am jaded by my experiences with AF, DA40LTD (two copies) and pentax bodies.

Right. Like I said, neither of us *really* have the data to definitely state what degree of misfocus is still within spec, or what percentage of cameras are out of spec, so we're just projecting our own experiences. Chances are, you are overly pessimistic, I'm overly optimistic, and the truth lies somewhere between. I still figure it's important for those who haven't already made up their minds to see both extremes of opinion.

It's always going to be an issue for someone with a body with no AF adjust. If you have a lense with a tiny bit of backfocus and another with a little front focus?

Here's another thing: I still don't really understand how a lens can be off. I mean, I can sort of see some things that might cause problems *sometimes*, if the initial focus position of the lens was really far from where it should be, so many turns of the screw are required to get it there. But consider, if the lens is already focused at the right place, there is nothing I can see the lens could do wrong that would cause a correctly-functioning camera to yield an out of focus picture - the camera would notice the picture was alreayd in focus, and not turn the screw. So even if the lens has something misadjusted such that it doesn't focus well when it starts far off, you should find each subsequent focus operation gets you closer and closer. And chances are you'd get both BF and FF (at different times, of course), depending on whether you started out focused too close or too far. Only if I saw symtoms like that would I be inclined to blame the lens. Do people who believe they have FF or BF problems in the lens actually see that kind of behavior?

Whereas I can *easily* understand how a camera could be off: if it's AF sensors are not *exactly* the same distance from the lens mount that the sensor is. It would strike me as almost miraculous if it were ever really each down to the micron. And if the Af sensors are misaligned, that would cause the phase detect system to simply be wrong about when an image is in focus, meaning no matter how manys you tried to focus, you'd always be off, by the same amount and in the same direction.

Anyhow, as a result of my thinking above, my assumption is that the vast majority of focus problems are caused by AF sensor misalignment - because that's an easily understandable cause of such problems, and the effect of that would be exactly what people typically report.

If you send it in to pentax do you think they will calibrate it all for you for free or tell you to deal with it? how long will you be without your gear for?

They do it for you for free if its under warranty. As for how long you are without your camera, I don't know, but given that I believe actual problems severe to consitute defects 9eg, unable to focus at f/2.8) are rare, I don't see why that should be a big concern. I mean, *any* defect might mean you have to send your camera in for service. Only if one believes that this particular defect is more common than any others should concerns about the actual repair procedure enter into it. Some small percentage of cameras are shipped with, say, flash units that are just DOA. I'm not saying it's a common problem, but it's happened. And it requires you to send the camera in for service. Unless you believe there are significantly more problems with AF than with dead flash unit, the fact that a camera "might" have an AF problem should be any more of a deterrent than the fact that it "might" have a dead flash, or "might" have a non-working shutter button, or "might" have a malfunctioning rear LCD, or "might" have any of the zillion problems a camera "might" have.

I'm not saying that a KX with a LTD lense won't be able to focus at f/2.8 but I am saying that in my experience that combo may need tweaking to be 100%

OK, then I think we've found that middle ground. Sure, no AF system in the history of photography has been *perfect*; there is no reason to expect to K-x to be perfect either. If you require perfection - or at least a higher degree of accuracy than has ever been possible before - by all means, get a more advanced camera tht provides this fine-tuning posibility. But if you're OK with a camera that is every bit as good as every other AF camera made over the last 30 years, the Kx should suffice.

[blu3ness]

Just got my K-X today, excuse my excitements.

I was a little surprised when I got my hands on the K-X, it felt a little small on my hand, however, I liked how my fingers are within reach of the controls very easily, having only 1 dial didn't seem to be a problem at all once I started using the camera, it felt very comfortable.

Second thing that got me worried was the "battery problem" that plagued some dpreview forum members, I was playing with the kit battery and suddenly it went to half-battery on the indicator. I turned off the camera, turned it back on after a while, back to full, so far no problems. I will do further testing tomorrow to make sure it's good..

I took a few pictures, since I only have the kit-lens (first DSLR, yay!), I can't really test to see if the camera focuses "a little soft" with a shallow DOP. I will test that when I get my 50 mm f/1.7 shipped to me.

I like it a lot so far tho, gonna go out and take lots of pictures tomorrow :P

[Class A]

I still don't really understand how a lens can be off.

So do you think that people who send back copies of a lens multiple times until they got one that doesn't FF or BF are crazy?

I don't know the full story either, but bear in mind that the AF system "sees" differently than the sensor. Remember that the stock focusing screen will "lie" to you when you choose apertures lower than f/2.8? I believe in a similar manner the AF system always looks at image parts at f/5.6.

Now consider a lens that has a focus shift between f/1.4 and f/5.6. This is a well know problem with fast lenses and stems from spherical abberrations. When the lens is wide open much of the image produced on the sensor comes from parts of the lens which are blacked out at f/5.6. Due to the spherical abberrations, most of the image produced at f/1.4 is in a slightly different focus plane than the one produced at f/5.6. This causes a focus shift.

If the AF system focuses at f/5.6 and the sensor "sees" the whole image at f/1.4, you'll get a mismatch.

As to why lenses will FF/BF at smaller apertures, I'm not entirely sure but I believe it has to do with the different ways in which the AF system and the sensor see the image (parts) and the fact that different lens copies have different imperfections.

[Marc Sabatella]

So do you think that people who send back copies of a lens multiple times until they got one that doesn't FF or BF are crazy?

No, although I suspect many of those people are simply mistaken - the problem was actually in their camera and they just *assumed* it was with the lens. I see people on these forums assume this all the time, without having actually checked. Or their reasoning is that the kit lens was fine, so the problem must be the new lens, but they are failing to realize that the larger DOF of the kit lens may have been masking problems with the camera. And of course, based on the number of pictures you posted where people are demonstrating their FF problem by shooting a bird in a tree and wondering why focus was on one of the branches instead, there's also quite a few people who simply return equipment that was actually functioning fine. And I put myself in that category - I once returned a DS because I thought it had an FF problem that turned out to be my own ineptitude in testing.

Anyhow, not saying that *all* problems are with the camera or with flawed testing methods, but yes, I do believe a "large" number of returned lenses *are* returned for invalid reasons. I suspect only a "small" number of lenses that are returned for focus problems actually have an issue. Of course, none of us has any proof one way of the other here, but I do admit that my opinion is based in part on the fact that I had yet to hear a good explanation of *how* a lens can cause focus problems. That is, what specific defect (like "misaligned AF sensor") would cause you to get OOF shots with a correctly-functioning camera? I'm sure there must be something, or lens manufacturers wouldn't ever take lenses back over focus issues, and you're hypothesis below is intriguing. The only explanations I've heard previously would lead to specific behaviors (like repeated attempts should gradually get better) that I almost never hear anyone describe in practice.

I don't know the full story either, but bear in mind that the AF system "sees" differently than the sensor. Remember that the stock focusing screen will "lie" to you when you choose apertures lower than f/2.8? I believe in a similar manner the AF system always looks at image parts at f/5.6.

Now consider a lens that has a focus shift between f/1.4 and f/5.6. This is a well know problem with fast lenses and stems from spherical abberrations. When the lens is wide open much of the image produced on the sensor comes from parts of the lens which are blacked out at f/5.6. Due to the spherical abberrations, most of the image produced at f/1.4 is in a slightly different focus plane than the one produced at f/5.6. This causes a focus shift.

Now that's a new one on me. Definitely seems plausible. But then I wonder, if this is caused by spherical aberrations, what can be done about it? That is, do we think maybe there is some mechanical coupling in the lens that is supposed to correct for this, and that could be what is out of adjustment? How would this work? And it also seems just not quite likely to me that the AF system would really be "seeing" f/5.6 the DOF is so large there compared to f/1.4, it seems it would be almost a miracle if it ever succeeded in placing the focus plane correctly (although I recognize that DOF might not mean the same thing to an AF sensor that it does to us). Furthermore, while the "focus shift" might be an issue when you *really* stop down the lens, any sense in which an AF system "sees" only f/5.6 obviously doesn't involve actually stopping down the lens, so I wonder how exactly it would actually affect the results. Intriguing thoughts, but lots of question marks remain in my mind here.

Anyhow, consider me a skeptic on this issue. I admit I have no proof as to what percentage of problems are with lens versus camera; I'm just stating what I suspect to be true.

[Class A]

Anyhow, not saying that *all* problems are with the camera or with flawed testing methods, but yes, I do believe a "large" number of returned lenses *are* returned for invalid reasons.

Quite possible.

But then I wonder, if this is caused by spherical aberrations, what can be done about it?

One thing that could be done about it is a correction by the camera. It knows your target aperture and can apply an AF adjustment accordingly. I have actually observed such a behaviour with my Tamron 18-250. The camera would refocus if I changed the target aperture (from say f/8 to f/3.5, but not from f/3.5 to f/8, though, presumably because the f/8 DOF is considered to cover the subject even though the focus is not optimally set on the subject).

And it also seems just not quite likely to me that the AF system would really be "seeing" f/5.6 the DOF is so large there compared to f/1.4, it seems it would be almost a miracle if it ever succeeded in placing the focus plane correctly (although I recognize that DOF might not mean the same thing to an AF sensor that it does to us).

I believe your last point applies. Even with a large DOF, there is still only one (theoretically infinitely thin) plane which gets rendered completely sharp. The rest is increasingly out of focus, it just happens to be acceptable blurriness (defined by a circle of confusion whose size depends on output size, assumed distance and acuity of the observer, etc.). I reckon a phase comparison system will still be able to focus precisely even though it sees with an effective aperture of f/5.6.

I admit to be speculating and don't feel great about that. On the other hand, most of what I say is based on facts, perhaps I'm not combining them entirely correctly. If you would like to obtain more certainty on the matter, I recommend calling falconeye to the rescue.

[kittykat46]

Based on my book knowledge of how the dSLR autofocus system worked, I had always believed that any true Front Focus/ Back Focus problem is body-related. After all, the Actual AF work is always done by the AF mechanism in the Body, the lens just reacts passively via the screw drive or SDM drive depending on the lens design.

When I really started testing my lenses, the surprise was that some of them were showing BF and some FF on the same body. They were all optically "Good" lenses, not malfunctioning, but just showing a bit of BF or FF. Logically, if its a body-based problem, I shouldn't be getting a combination of FF and BF with different lenses....so there is some other factor at work here.

On top of that, my previous chat with camera technicians who have carried out such adjustments before, shows that there is definitely a need for individual lens-based focus adjustment, on top of overall body-based adjustment.

The other obvious fact they pointed out is the majority of cases of users returning lenses because of reported FF/BF Focus problems are not "True" - either user error, missed AF, body setting problems, etc.

Well, the technicians carry out the AF adjustments without a deep understanding of the optics involved. I guess we'll need information from the actual design engineers to understand what is the real technical issue surrounding FF or BF problems.

[Ben_Edict]

Me too... the quality (or lack of) in the accuracy of the fast glass is pointless with no AF adjustment. What's the point of paying for 2.4 / 2.8 if you ALWAYS haqve to shoot at 5.6 to get a sharp image. Knowing that some of my glass is 'off' makes AF adjustment a must.

Does the KX have it?

I think, that generally the expectations are way too high. The K-x is an entry level camera. 95% of buyers will live with the kit lens and perhaps add a 50-200 and that's about it. Using fast glass on an entry level camera will be the rare exception and it is only a statement to the advanced state of photogs here in this forum, that this topic even shows up.

Also, I don't think the things discussed here as "Missing", like the illuminated AF-indicators or the AF-adjustment, need top be present in an entry level camera.
Ben

[Iskandar Taib]

1200 shots off AA's?

what are these mythical beings?

Don't know about 1200, but I was getting between 800 and 900 with my K100D Super.

Iskandar

[Iskandar Taib]

When I really started testing my lenses, the surprise was that some of them were showing BF and some FF on the same body. They were all optically "Good" lenses, not malfunctioning, but just showing a bit of BF or FF. Logically, if its a body-based problem, I shouldn't be getting a combination of FF and BF with different lenses....so there is some other factor at work here.

On top of that, my previous chat with camera technicians who have carried out such adjustments before, shows that there is definitely a need for individual lens-based focus adjustment, on top of overall body-based adjustment.

From what I've heard, this is indeed what exists - there's some "secret menu" you can get into (the steps to get in have been "outed" for the K10D, but not for the others) where you can, among other things, adjust the individual correct focus for each and every lens you have. The lenses have chips in them that lets the camera tell what lens is attached.

Never done it myself, mind you. And this is all based on hearsay, so no flames, please. Correct me if this is wrong.

Iskandar

[Aristophanes]

I think, that generally the expectations are way too high. The K-x is an entry level camera. 95% of buyers will live with the kit lens and perhaps add a 50-200 and that's about it. Using fast glass on an entry level camera will be the rare exception and it is only a statement to the advanced state of photogs here in this forum, that this topic even shows up.

Also, I don't think the things discussed here as "Missing", like the illuminated AF-indicators or the AF-adjustment, need top be present in an entry level camera.
Ben

Depends on what the competitors are doing, not your opinion. If they do it and you do not (and they are doing it), you lose sales and market share.

Also, generalizing about what "95%" of buyers will and will not do is silly and a sales-killer. Pentax absolutely MUST have more than 5% of K-x buyers buy up and add lenses. Margins on lenses are probably around 20%, whereas on bodies, less than half that. That's the whole point of mid-range glass like the 16-45 and 17-70, not to mention the bulk of the third party market. However, there's no body mid-range from Pentax. It's like a great, big hole.

Not to mention that both Pentax Canada and USA do NOT authorize selling the K-x body without the kit lens. Any item sold as such is officially grey market. Compared to Nikon especially, Pentax is very rigid. That, too, kills sales. People like options.

[Eruditass]

Agree with that - many people are unhappy they can't swap the 18-55 out for a 16-45 or 17-70 as an option

As for grey market - does that mean they won't warranty places like Adorama that sell body only options or something?