Why is Quality Control so low?

Discussion in 'Canon EOS' started by ted_holm|1, Apr 10, 2011.

  1. Why do we spend thousands of dollars on photography equipment and put up with low quality control?
    Look at lens reviews, even the ones done by DXO labs, claims to find bad copies of lens. When I look at customer reviews of lens, it is not uncommon to read that a person has mailed back one or two lens to find on that works.
    Digital cameras now have micro adjustments to have the camera owner makes adjustments to lens that are not calibrated correctly.
    If I buy a lens for a Nikon or Canon camera I would think that this new lens it is going to work fine out of the box.
    Anyone else upset with buying a lens and hoping it is a "good" copy. And what happens to the "bad" copies that go back? Are they resold to the next customer?
     
  2. Ah dunno about you, but I've noted a gradual increase in optical quality during the past 40 years of shooting. Nothing is perfect but my 13 x 19 (on Epson) and 24 x 36 (lab) prints have never looked cleaner or sharper from small format cameras like the 5D2 and 7D. My old Nikon gear from the 60s and 70s was decent but not in the same league as my L optics (still have some of those ancient AI Nikkors).
    Sure there are a few lemons out there but most QC problem are in reality due to poor photographic skill and even worse testing methodology. These fools henceforth blow BS out their arses on forums, making it seem like a bigger problem than it actually is.
    Although I have MA on my 5D2 and 7D, none of my 20 or so lenses need adjustment so I returned it to defaults after monkeying with it for a couple weeks.
     
  3. Because a very small percentage of the people with fine working tools will post that their stuff works.
    It's the ones with the bad copies that complain.
    This is an effect of the "global village" in which we live.
     
  4. I have never had a quality issue with any Nikon equipment, but I have from Sigma. I should imagine the QC of Nikon's consumer equipment would not be as rigorous as their pro equipment, as a necessity to keep prices low (relatively), but I hear of very few issues. I would expect Canon's QC to be of an equally high standard, especially for items costing thousands of dollars, as you put it.
    I completely agree with the replies above.
    BTW, I came here from your Nikon thread, which has been closed, and members redirected here.
     
  5. I have not knowingly had a problem with my lenses yet. In fact, it has created a paranoia that I am not good enough to realise they are poor copies.
    I am convinced that most reports of poor copies are due to either user error or pixel peeping to levels that are nearly irrelevant in the real world. Any manufactured product has a range of quality output where 'stellar' is as much a freak as 'substandard' and one 'disdvantage' of the internet is that it can create unrealistic expectations - someone posts about a stellar copy of a lens bt everyone automatically assumes they should all get that copy.
    I was not a forum poster until 2006 but did people complain in such volume during the days of 35mm film - I suspect not and surely the manufacturing quality has improd over that time?
     
  6. I was not a forum poster until 2006 but did people complain in such volume during the days of 35mm film - I suspect not and surely the manufacturing quality has improd over that time?​
    Optical design and manufacturing are very mature but have been slowly evolving for generations. I dare say there is little significant difference from 2006 to 2011. Toss in a couple decades and you'll see significant gains in design. Realize some of the finest optical designs from the 80s and earlier still hold their own against our latest and greatest. The EF 200 1.8L and 300 2.8L come to mind...
    During the film era most folks printed 4x6 and were happy. A few hardcore made large prints, projected for viewed with loupes on light tables. Today any fool can zoom to pixel level and beyond, quickly crying wolf when their one-handed stab technique yields mush.
     
  7. If you think the QC is low, just take a moment to read the repair manual for the Nikon 105mm VR macro lens (and I'm sure the equivalent Canon lenses). Adjustments are made in the reassembly of the lens that are really incredibly precise (and calibrating the VR looks like a real task). In fact roughly 30% of the repair manual is disassembly/reassembly, and all the rest is quality control and adjustment guidance.
     
  8. Ted, your question would make more sense if worded, "Why do we spend thousands of dollars on photography equipment and put up with poor images from not learning to use it correctly?"
    After too many bodies, and even more lenses, I can't say quality control has ever been an issue for me. I'm like Mike Hitchen, maybe one day I'll be good enough to recognize my bodies and lenses aren't up to snuff.
     
  9. like girl friends I try to make the lens work with what I get, I guess I ain't to picky, or just lucky, at least with lens as girl friends tend to come with more problems so they get sent back, still have not had to do that with a lens, two cameras yes
     
  10. I have said it before and I will say it again: For optimum performance a lens need to be calibrated to your body (especially telephoto lenses). If it is not properly adjusted, you may have issues with sharpness and/or AF accuracy.
    Then again, AF accuracy is not that perfect as we want it to, especially on non-highest end bodies. Lots of practice will give you the experience to overcome the drawbacks of the AF system most of time -- so always refocus and shoot twice. Lenses with consistent focus errors are very rare and usually linked to third-party manufacturers like Sigma.
    I myself have owned lots of Canon EF lenses -- all bought on the fickle used market -- and have never had a "bad copy". If you feel insecure, bring your lens and body to the Canon repair service and have them properly calibrated to each other. I considers this necessary maintenance and not a quality issue, because there are differences in the high-end and low end bodies that can have an impact.
     
  11. let's assume for a second that Canon (and Nikon) make millions of lenses each year...not a bad assumption. Of those say 1% are defective or have issues. That would be 10,000 bad lenses per year.
    Some of those will be returned - where they will be refurbished and resold. Some returned and trashed - and others not returned at all - just held on to and suffered with and gripped about on PNet and other forums. And because Pnet and others don't verify identity - that person could have 50 accounts here - and complain / agree 50 times.
    Most of us - if we get a lemon out of the box - will return it promptly. Others will try to sell it on Flea - Bay or somewhere similar. Then the "bad" lens is passed on to others and they complain too...
    Of course there always is that other category - the folks that buy something and think it is going to solve all the world's problems - and that you will be able to take a photo like is shown the ads every time - without really trying.... That group won't be happy with anything.
    Dave
     
  12. It's the ones with the bad copies that complain.​


    There's the problem. You don't want to be using copies, get an original!
     
  13. Cameras and lenses of today are far more complex than the ones we grew up with. There are many shooters that never used film and manual lenses. They expect a 300mm AF lens to work like a P&S and they whine when it doesn't.
     
  14. When I look at customer reviews of lens, it is not uncommon to read that a person has mailed back one or two lens to find on that works.
    High end professionals have long had a practice of regularly try several lenses of the same mode lens before picking the best of them. This process long predates autofocus or digital photographic technology and for that matter the internet. Cinematographers do it too. The earliest example I have heard of this practice is W. Eugene Smith doing this back in the 1950s.
    Digital cameras now have micro adjustments to have the camera owner makes adjustments to lens that are not calibrated correctly.
    The function of micro AF adjustment is to fine tune individual lens and individual body AF performance to each other, to fine tune and specific system, not to correct for lenses that are not calibrated correctly.
    Your screed leaves out the possibility that user error is a factor, which I suspect is is a factor in 90% of the cases when someone says they bought "a bad lens".
     
  15. Most of the "faults" are user error. DXO couldn't make a meaningful test of tissue paper if they had a cold.
    Read this to understand why so many people think they have a "bad" lens.
     
  16. The question is based on a series of false premises. Posting it twice makes it no better.
    Quality control has generally improved, not deteriorated. Unfortunately the ability of some people to learn how to use complex machinery has NOT improved, and is unlikely ever to do so.
     
  17. The internet has a way of purifying, enhancing, and then rapidly breeding ill-conceived, poorly informed, and frequently malicious memes about bad lenses. It's like a petri dish made especially to support the growth of threads, comments, tweets and blogs that include the food-like phrase "bad sample" and which do not include the powerful antibiotic phrases "rigor," "method," "experience," or "used a tripod" in the same paragraphs.

    There has never been a better time to buy photographic equipment of any type.

    Also, Ellis gets the gold star for correctly using today's secret word: "Screed"
     
  18. I have used probably over 50 F mount lenses on my Nikons (by Nikon and Zeiss) and never have I run into a "bad copy". I think that's very telling. I'm very demanding on my lenses. Bad copies of good lenses are very rare.
     
  19. I don't think that your perception of low quality control is really accurate. I agree with those who have already pointed out that the original post is based on a series of false premises.
    One of the problems with the increasing ability to post immediately about anything and everything and then read what everyone else has written in this manner is that certain false memes gain traction and appear to be more real than they actually are. In reality, the quality control on modern photographic products is excellent, as good or much better than it has ever been. And the products themselves are better and more powerful as well.
    Yet people, whether or not they are actually qualified to do so, can look closely at some 100% magnification crops on their computers and claim that product X is out of focus or distorted or whatever. First of all, quite a few of these claims attribute the perceived "flaw" to the wrong thing. To offer one example, we periodically read that "previous model X with 12MP produced sharper images that new model Y with 21MP"... when someone looks at both at 100% and fails to notice the obvious fact that he/she is looking much more closely at a smaller area of the frame in the second case. Or someone inspects the same way and complains about "excessive CA," when a) CA has always been present in similar lenses, b) it is probably better today, c) it is not going to be visible in real world photographic images, and d) it is easily corrected in post. There are many, many more examples of this.
    Even when we know that these sorts of claims are misguided and/or just plain wrong, they gain traction on the net. People, especially some who want to find a conspiracy behind every human endeavor, load these false extrapolations up with the baggage of "someone is trying to pull a fast one on us" or "these companies are cutting corners" and so forth. And the natural instinct to look for explanations that involve unethical behavior by others kicks in... and we get claims like the one that started this thread.
    Finally, for some reason photography seems to attract, among its many adherents, a certain crowd that looks for a mystical technical perfection to provide satisfaction. In some cases it seems that photographs are not really all that important, and that even tools of photography are only important to the extent that they satisfy this desire for technical perfection. I have some news. Photography is about photographs, not about equipment. No technology is perfect. Brilliant photography has been and will continue to be produced with the gear that exists.
    Dan
     
  20. “Why is Quality Control so low?”
    I cannot speak for all industries but the quality control programs I have witnessed in the photographic, automotive, electronic, and food industries are excellent. How did you arrive at your conclusion?
     
  21. I've probably bought (and sold...) more Canon EOS gear then the average Canon user. I've never gotten a lens or camera that was defective out of the box. I've had one or two lenses that needed minor repair after a few years of use, but that's pretty much to be expected. Their QC seems pretty good based on my experience.
    In cases where quality control is low, the reason is cost. If consumers demand the absolute lowest rock bottom price on gear, one way to lower production cost is to cut back on quality control. Instead of testing every camara or lens, you test 1 out of every 10 or 1 out of every 100 and instead of testing every function for perfection, you test a few functions for "within spec".
     
  22. One bad apple on the tree does not constitute the need for a chainsaw. Just pick another one.
     
  23. Ted Holm , Apr 11, 2011; 02:54 a.m.
    said...
    Why do we spend thousands of dollars on photography equipment and put up with low quality control?
    Look at lens reviews, even the ones done by DXO labs, claims to find bad copies of lens. When I look at customer reviews of lens, it is not uncommon to read that a person has mailed back one or two lens to find on that works.
    but,
    Puppy Face [​IMG], Apr 11, 2011; 03:19 a.m.
    said...
    Ah dunno about you, but I've noted a gradual increase in optical quality during the past 40 years of shooting.
    *******************************************************************
    I think both statements have merit. The causes of a perception of low quality could be from:
    1. More product. If a name brand lens is going to have a dud at the ratio of 1:1000 and they sell 10,000 lenses they are going to have 10 problem lenses. If they sell 1,000,000 lenses they are going to have 1000 problems.
    2. As workforces mature, less emphasis is put on "doing it right". Look what happened to Detroit automobiles years ago, talk about low quality! And to think that their fathers used to make Duesenbergs and Packards. Japan's workforce culture has matured and the same type of social changes are happening. Have you read Consumer Reports for over 50 years? First the American Cadillac had the best repair record (even though the somewhat socialist Consumer Reports didn't seem to like the idea of folks owning Cadillacs). Some years later it was the German Volkswagen and Mercedes. In a later time period it became Japan's Honda, Toyota and Datsun. Now the Korean Hyundai is getting high marks. That is a good example of changing workplace ethics and customs.
    3. Corporate Directors elected by fund managers have no incentive to build a company's reputation for the long term. Their fund bosses just trade stocks in and out, They only want a good financial quarter, this quarter. *X&$#**X&$#**X&$#**X&$#* rolls down hill. The worker and his supervisor know well what the current crop of temporary masters want. Only those with enough protective seniority can put in the effort to make a good product and... only if they want to.
    The perception that lenses are better may be partially due to:
    1. Advanced science over the last 100 years.
    2. Computer controlled grinding which makes the batch to batch differences smaller.
    3. The use of electronics for the controls which can have tighter specs.
    4. Better materials.
    Yes, I think there is merit on both sides of the argument.
     
  24. >>> Why is Quality Control so low?

    That hasn't been my experience.
     
  25. I stand corrected. I respect each one of the people the folks that have taken time to share their opinions.
    Many people have taken up the art of photography and find that today’s equipment can be overwhelming.
    I have been using a Nikon lens and am in the market for a Canon camera. I have been looking for a wide-angle lens, reading reviews on several lenses. Then I read a review on the Tokin 14-28 FX pro lens. I was shocked to say the least.
    Back when I was shooting film, we didn’t need to “fix” our photos in Photoshop or lightroom. We could use all of the f stops on our lens and not worry about softness, we would only worry about depth of feld. No sweet spots. Today we seem to shoot a lot more photos hoping that a few are winners.
    So I am waving the white flag, thank you for all your input.
    Happy shooting to everyone.
     
  26. That was a fine wrap up, Ted. I think quite a few people would have laid low after such a barrage.
    [start rant]
    By the way, reading some remarks about rigorous testing... I shoot all my pictures hand held, I tend to pixel peep and I still think that a lot of the shots are sharp. Maybe my standards are lower than others, maybe my hands are more steady than others. I don't know but I do know that for me -unless I'm pushing it- the results are sharp enough.
    [end rant]
     
  27. Now, Ted, you sure didn't just wave the white flag. But you probably are looking at back-in-the-day through some rose colored ND filters.

    Back when you were shooting film, you "fixed" things in advance by matching film stock to light source, and then relied on an experienced pro in the lab if prints had to be critical (enter considerable work in the darkroom, which has simply moved to the desktop, instead). You can still leave all of that to a pro, and just shoot - just like you did with film, if that's how you like it.

    And of course lenses have always had sweet spots. Especially at the consumer price level. They're better now than they used to be. Your concern about "using all of the f stops and not worrying about softness" is related to diffraction. That's a laws of physics issue, and has absolutely nothing to do with quality control. Diffraction has always been an issue, and always will be. Its impact varies with sensor/film/body format and which lens is in play. Some people shooting with smaller format digital bodies seem to think that diffraction is some sort of new-on-the-scene curse that came along with digital devices or newer lenses, and that what happens at f/22 on an APS-C format sensor is somehow the result of sloppy engineering at the factory. It's the photons, man - they are what they are, and they do what they do when you chase them through a small hole to project them onto a small surface.

    As for what "we" seem to do, in terms of shooting more and hoping ... how would you define "we?" People who are critically concerned about results, and who know enough to fuss over issues like diffraction, are as deliberate today as they were in the past - possibly more so, since the photographer is the lab these days in many cases, and can see things at the pixel level the moment they happen. The difference is that now they can experiment more because they don't have to involve a wet lab and the wait time to see how a different technique is panning out.
     
  28. I think a lot of the qa/qc issues are operator error, you know, RTFM. I've got about 7 functioning eos bodies bet. digital and film, and I'm not about to send them to Canon to be calibrated. If I want critical focus I use a tripod and mf.
    I've been shooting since 1980 and buying from B&H almost the entire time. I've only had one thing defective out of the box, a 5d-1 body, but it was repaired by Canon. I seriously doubt it underwent a QA/QC check at the factory that included firing the camera and recording to a CF card, because if it had my body would never have left the factory.
     
  29. Why is quality control so low? Because only one guy works there.

    Get it? One guy? Solo?

    Bwaah-ha-ha-ha!

    All of my Canon stuff has worked well so far. My Nissan is another story.
     
  30. There was a great post on the Nikon forum a couple of weeks ago. The poster was complaining because they thought that their new D7000 was supposed to come "with HDR.". They were unhappy that HDR wasn't a built in
    feature. Unrealistic expectations.
     
  31. We could use all of the f stops on our lens and not worry about softness, we would only worry about depth of feld. No sweet spots.​
    Actually, all those things were there, but in the absence of the internet, no one much noticed them.
    In actual historical fact, the photomagazines back into the 30s talk about professionals going to the store and trying out one lens after another of a particular model to find the "perfect one". It was certainly more justified then than now.
     
  32. If you feel insecure, bring your lens and body to the Canon repair service and have them properly calibrated to each other.​
    I wasn't insecure or paranoid:
    • My brand new EOS3 did play up consistently whenever it was cold and/or a bit damp (like early morning chill, conditions that never seemed to affect my backup Elan 7e body), namely the display inside the viewfinder didn't work, but the body came back from Canon's own service with "no problem detected" note -- it was neither detected nor fixed in spite of my detailed description of when and how the issue was manifesting itself.
    • When the flash started to short-circuit the batteries, they blamed it on the batteries (which I changed when I noticed how hot they got, but the fresh set didn't fare well either, which led me to believe the Speedlite was at fault) and send it back to me without even touching it. Repairing it "independently" wasn't economical -- would have cost me more than buying a new one -- so I had to toss it out.
    • When 28-135 IS came back from a repair (I dropped it, or rather a gust of wind blew my tripod over, causing the aperture to get stuck), I had to send it back when I noticed a loose screw rattling inside the lens (I could see it through the front element, and it was sizable, not one of those super-tiny ones), and when I got it back the second time, the zoom action was never smooth again (persistent squeaking) and IS was making loud noises it didn't used to, but they refused to replace the lens although they were clearly unable to fix it properly.
    So yeah, I'm with Ted.
    Not just poor manufacturing QC but also unacceptable Customer Service x(
    P.S.
    Oh, that's a good story too. When the collars on my Gitzo tripod got stuck, upon inspection Gitzo service advised me they would have to cut and replace them, and quoted me a price to which I said thanks but no thanks. I then took it to some little workshop, and guess what? -- not only did the guys there somehow manage to get the collars un-stuck, but they didn't even charge me for their service!
     
  33. My images have generally improved since I moved from shooting mainly with Canon FD gear to shooting mainly with EOS film and digital gear. I attribute this to four factors: improved metering, autofocus, (marginally) improved optics, and my own improving technique.
    I have never thought that Canon had "quality control issues," then or now.
     
  34. Ted, i have a story that might make some sense. I have been shooting my Mamiya RZ 67 for years using film and scanning and never once did i stop to think of quality until i purchased a digital back to use with it. At first I thought the back had issues because i was getting some pretty crappy images and even thought it might be a calibration issue. As a matter of fact i even got word from a reputable camera repair store that i had to send them both the camera and back for calibration. After lots of trial and error i figured out that my problem was with with me and not the back at all. I now shoot and get excellent results with the digital back. In fact i get better results than from my scanned film even when i scan with drum. I know i'll get some slack from this comment.
    Ed
     
  35. It's always easy (too easy in my view) to blame "pilot error", but consider this. Reports of bad lenses are not rare. For example, "I bought the Whizbang 85mm f/1.0 and it didn't focus properly so I had to return it for a replacement." If it is an unsolicited comment by a random poster then a high rate of reported defects could be explainable by selective reporting, as has already been pointed out. In other words, maybe there were only a handful of bad Whizbangs made, but if someone got one they are more likely to make a post on it than the satisfied users.
    However, if it appears in a review in which the author started out with the intent of reviewing the Whizbang then a bad report is much less likely to be selective reporting. In that case a bad report is much more likely to represent a non-rare event.
    What is more telling is that some persons have reported having to make multiple exchanges before receiving a Whizbang that worked well. This is telling on two accounts. First, if the author eventually obtained a good Whizbang it seems more likely that pilot error is not at fault, but rather the lens (or whatever product) itself was at fault.
    Second, based on probability theory, and making the assumption that sampling of independent units is statistically independent, if someone receives two bad Whizbangs in a row it is likely that the defect rate is high. For example, if the defect rate is 1% then the chance of receiving two faulty Whizbangs in a row is 1% squared, or 1 in 10,000. Thus, two consecutive Whizbangs would be a rare event, and reports of consecutive defective units is a pretty good indicator that the defect rate is high. Continuing this line of thought, if the defect rate is 1% then the probability of receiving three defective Whizbangs in a row is one in a million, a very rare event. Therefore, if anyone is receiving two bad Whizbangs in a row it would be reasonably strong evidence for poor quality control, and three bad Whizbangs in a row would be very strong evidence of poor quality control.
    Now, let us consider quality control theory. The current standard for best practice at the best companies is something known as six sigma, which basically means that the defect rate is approximately one in a million tries. What constitutes a "try" could be open to some interpretation. Does that refer to the individual parts, or to the assembled whole? Nevertheless, based on the not-infrequent reports of lens defects it seems to me that the camera companies today are probably not operating at a six sigma level.
    Finally, it is probably worth addressing the question of just what would be an acceptable defect rate. Is 10% unacceptable? It would be in my view, though others may disagree. Is 1% unacceptable? In my view that would also be too high. In my opinion a defect rate on something like a camera lens should be 0.1% or less. I say this from the point of view of a consumer, but also someone who knows a little about quality control. (I am involved somewhat in a quality control program at my place of employment, though what we deal in is not similar to optical components.)
    Further discussion is welcome. Please throw flames if you want, but please keep it civil.
     
  36. Alan,
    Or you could read the link that I posted many comments ago that explains exactly why people think they need to go through 4 lenses before they get a good one. Notice almost all of these "quality control problems" are sharpness related, very rarely anything else, doesn't that seem strange to you?
    To make it easier to find the common sense within this thread here is the link again.
    Civil enough? :)
     
  37. Ted:
    I own more than $30K of Nikon equipment, most of it pro-grade. Fully one-half of it has either been repaired by Nikon under warranty, or gone back immediately to the vendor on purchase due to gross defects.
    Nikon, Canon and the others have outsourced all quality control to their customers, because they can. Most amateurs cannot tell a badly decentered lens from a good lens. Decentering doesn't show up on pictures of your cat.
    Your question could be rephrased: why was Detroit able to foist two decades of crappy cars on the American public? Because they could.
    My policy is to buy, test quickly, and return defects aggressively. Every time I return a defective lens to B&H or Adorama, I know that it will be repackaged and resold as new to someone who doesn't know any better. At first, I felt guilty.
     
  38. Because of the common people notion that everything has to be C H E A P! If we would spend more money on more quality equipment, then the quality would be higher all way through. Im going Leica the next time I switch.
     
  39. Now, let us consider quality control theory. The current standard for best practice at the best companies is something known as six sigma, which basically means that the defect rate is approximately one in a million tries. What constitutes a "try" could be open to some interpretation. Does that refer to the individual parts, or to the assembled whole? Nevertheless, based on the not-infrequent reports of lens defects it seems to me that the camera companies today are probably not operating at a six sigma level.​
    Six sigma and 'defect' rate is only relevent in the context of what the company defines as an acceptable quality, and the level they define as acceptable will be governed largely by the market sector they are aiming for. High volume sales at low price will need lower production standards, lower levels of accetpability and lower rejection rates than a premium brand.
    The problem with lenses is that they can only operate in conjunction with a camera. Suppose you have what you define as a 'stellar' lens which means you have a body to which it is (coincidentally) well-matched. You upgrade the body and the lens/body combination suddenly becomes a real dog - which do you blame? My guess is you would automatically blame the body because it is a new component despite the fact that both are within manufaturing acceptability.
     
  40. I own more than $30K of Nikon equipment, most of it pro-grade. Fully one-half of it has either been repaired by Nikon under warranty, or gone back immediately to the vendor on purchase due to gross defects.
    That's just amazing. I've never purchased a Nikon product which had an obvious defect that affected real-world results in any way. And I've spent much more than $30K. I did buy an F5 which has a faint scratch in the metal rails on the film path but it doesn't affect the performance of the camera in any way, and I got it at 20% off the normal price. Could you list examples of specific equipment and what was the "gross defect" that you found in it?
    Every time I return a defective lens to B&H or Adorama, I know that it will be repackaged and resold as new to someone who doesn't know any better.
    How do you know that? I would think that re-selling used equipment as new is against the law. If it's not then it's a case of inadequate consumer protection laws in your country. At least around here, you cannot return an item which is not faulty, and if it was not purchased mail-order. If it is mail-order and not faulty, then it needs to be an unopened box. Otherwise they have to discount it and sell it as not new.
    If I don't like an item I bought, I will trade it for something else, but I take the usual 30% hit. It's my fault for not doing my research properly and buying something that isn't good enough for my needs, and I accept responsibility for that.
     
  41. david_henderson

    david_henderson www.photography001.com

    Maybe Canon and Nikon should make the same equipment available at say three different price points. The first as now with their preferred level and cost of quality assurance. The second with a demonstrably more aggressive level of QA meeting "Six Sigma" (or whatever) level of testing at whatever cost/price that requires. Then with no or perfunctory post- manufacture QA at a lower price than now.
    It'd be interesting to see what the price differences would be in the context that you'd expect warranty costs to vary inversely to the amount of quality assurance input, and it would be interesting to see what alternative people actually voted for.
     
  42. Most amateurs cannot tell a badly decentered lens from a good lens.
    If a lens has for example, a soft left 1/3rd of the image, it would show up in real-world use, and most people will notice it. This was the case with my 20mm Nikkor but the problem with that was not quality control, as on film the whole image was sharp; it was the case of a design which was optically incompatible with today's digital cameras. The slight asymmetry was adjusted by Nikon so that it was now both sides which were slightly unsharp, and after a month of almost no use, the lens was back to the its old asymmetry. There was nothing that could be done since the lens was not built robustly enough to hold the alignment so well.
    But with lenses that are used with their intended cameras there is rarely a problem. In another thread, you mention a decentered 24/1.4. Now, did you point the lens towards a flat surface, focus and then investigate the whole image ? If you did, this is incorrect use. A lens like this is not flat field and there is no guarantee of the shape of that field. What you need to do is live view manual focus the lens on the exact spot that you will investigate in the image. And then focus on the other side, and take another image. If one side cannot be focused to yield good definition, on a specific spot, and the symmetrically opposite side can be, then you probably have a problem. (And do several repetitions of the manual focus operation since the turn of the ring is so quick.) But if you just focus on one spot and assume the focused plane is a plane, and that you should get sharp detail across the whole image you're making too many assumptions about the design of the lens. Autofocus lenses are designed to be sharp at the point under the selected focus sensor, and there are no guarantees of other places. Also, it happens to be that most Nikkors are optimized for relatively short distances and if you, like Lloyd Chambers seems to like to do, expect it to focus precisely on a subject that is ten meters from the camera or further, without confusing between details that are further behind e.g. the branches you're focused on, again this is a misguided assumption on how the AF system and the lens work. For distant subjects if you want the best results you must focus using live view. And field curvature can be very strong at long distances with wide angle Nikkors. Nikon designed the f/1.4 wide angles for typical PJ shooting and they (similarly to the f/2.8 zooms) give best results around something like a distance of 2m to subject. You can get good results at long distances but you need to forsake autofocus and be very careful about how you do the test before coming to conclusions about the optical quality or quality control of the lens. If you just AF on a subject 30m from the camera there is very high chance that the AF will focus on something that you don't intend, and the precision of the system is insufficient at these distances in any case.
    In real-world use, in the hands of an experienced operator, the f/1.4 AF-S Nikkor wide angles yield superb results. I just wish Nikon had come out with them a few years earlier.
     
  43. William Michael

    William Michael Moderator Staff Member

    "Every time I return a defective lens to B&H or Adorama, I know that it will be repackaged and resold as new to someone who doesn't know any better."​
    I find this one of the MOST interesting comments I have read in this forum.
    I don’t live in the USA and I have often commented upon the capacity to return an item simply because of change of mind; but this is different.
    If I understand correctly this is stating that faulty product is returned to the Retailer and subsequently it is repackaged and resold as a new item as if directly sent from the Manufacturer.
    There are two questions I would like answered:
    1. How do you, (James), know this is the practice. (Evidence?)
    2. Is this practice legal under whatever the USA equivalent is of laws governing “Trade Practices”?
    WW
     
  44. ShunCheung

    ShunCheung Administrator

    B&H and Adorama have to be extremely stupid to repackage defective items as new and sell them again.
    Last month I bought a refurbished Nikon zoom lens from B&H. Upon receiving it, I realized that it has some sharpness issues on the long end. Most likely the vibration reduction mechanism is defective since I could see a little motion blur shooting a still subject from a tripod with VR off. I called B&H, and they paid for the return shipping and gave me a full refund. In other words, B&H did not make any money on that order; in fact, they lost money because they paid for return shipping plus all the paperwork and labor for shipping.
    Whether it is legal to do so aside, if B&H ships that same lens to someone one, most likely the next customer will discover the same problem and return it. Losing money once on a bad lens is unfortunate. Why would B&H be so dumb to keep losing money on that same lens over and over? Someone so stupid will not stay in business for long.
     
  45. Here's my take on QC, based on personal experience.


    For many years, my philosophy ranged consistently on the optimistic side, like most posters here, to the point that I never bothered
    testing the gear I purchased (mostly second hand).


    Then, one day, I bought a sigma 50/1.4 ef mount, new. Went on a trip without testing it. Looking at the pictures back home, I see
    something consistently wrong : a significant number of them are severely back focused. Others are tack sharp. I blame it on my
    mistakes. I continue using the lens, with a high percentage of misfocused shots, much higher than with any other lens.


    I finally decide to run a serious test. What I find out is that the lens focuses well at short distances, but vastly backfocuses past 3 or 4
    meters. I send it to Sigma France. Comes back with absolutely no change. I call them. They explain that they test focus at a distance
    of 1.5 meter, and they found no flaw. I have to describe the issue and offer to send sample files. I send the lens back, they
    acknowledge the issue, alledgedly fix it, but when i get the lens back it still backfocuses in the same way. Another call to sigma in
    France : they offer to send the lens to Japan. Two months later, the lens is back, and now works properly.


    My conclusion : i should have tested the lens out of the box. Because I began to doubt the lens instead of my skills ten months after
    the purchase, the lens had gotten enough use that it could no longer be exchanged (usage marks). Which is why I lost so much time
    wth the lens being sent back and forth. A word has to be said for the sigma people, who were very dedicated and helpful.


    From now on, i will test any item i purchase. I've learned my lesson : *X&$#**X&$#**X&$#**X&$#* happens !
     
  46. Lupo Lobo: As this day moves on I hope that Henry Posner from B&H or the representative from Adorama can shed some light on what actually happens with returned merchandise.​
    Simple -- a defective product is returned to the source from whence it came. Reselling a broken item hoping that customer #2 won't stumble upon the defect which caused customer #1 to send it back is not merely ludicrous; it's retail suicide.
    Matt Laur: The internet has a way of purifying, enhancing, and then rapidly breeding ill-conceived, poorly informed, and frequently malicious memes about bad lenses.​
    Seems to me the internet has a way of purifying, enhancing, and then rapidly breeding ill-conceived, poorly informed, and frequently malicious memes about plenty of things...
    James Meketa: Every time I return a defective lens to B&H or Adorama, I know that it will be repackaged and resold as new to someone who doesn't know any better.​
    Far be it form me to speak for Adorama, but as far as B&H is concerned this is a baseless and despicable calumny. James, it's my personal opinion that what you think you know and the facts reside in two different time zones.
    William W: How do you, (James), know this is the practice. (Evidence?)​
    I wouldn't mind reading the answer to this too.
    Henry Posner
    B&H Photo-Video
     
  47. William Michael

    William Michael Moderator Staff Member

    "Why would B&H be so dumb to keep losing money on that same lens over and over? Someone so stupid will not stay in business for long."​
    Umm: that's what I thought, too.
    That's why I asked the two quite specific questions.
    Also that's why I think James' definitive assertions are so very interesting.

    WW
     
  48. William Michael

    William Michael Moderator Staff Member

    Mr Henry Posner,
    Thank you for a speedy and direct response.
    I await with much interest, Mr James Meketa's response to my request for evidence, to his previous statements.
    WW
     
  49. Lupo / James
    I can't speak for all other retailers, but at Adorama, faulty items are returned to the manufacturer or distributor.

    That doesn't mean that there aren't occasional problems; the manufacturers and distributors, who supply to Adorama, and of course, other retailers, do allow us to return equipment as new, because of over-stocks, exchanges etc. This can be a particular issue after the Christmas holiday when we all need to balance out our inventories.

    Even if we at Adorama are completely thorough when packaging up these items for return, there is no doubt that less scrupulous retailers could well be receiving returns, trusting the customer's explanation that they never opened the box - packaging them up for return to the manufacturer or distributor. These items then form part of the inventory for sending out to other retailers.
    No retailer can inspect the contents of every packaged received into the warehouse.

    If the manufacturer doesn't check the equipment before sending it back out, from the customer's perspective it is the retailer who looks bad, even though it may well have not been their fault.
    At Adorama we do have liberal return policies and there is no doubt that on occasion this is taken advantage of - although this does not exempt us from more closely checking any and all returns, there is little doubt that some retailers don't.

    The problem is that we are all relying upon humans to carry out the required procedures at every point in the transaction, which can, and does, result in mistakes at any place along the line.

    Helen Oster
    Adorama Camera Customer Service Ambassador

    helen@adorama.com
    http://twitter.com/HelenOster
     
  50. William Michael

    William Michael Moderator Staff Member

    WOW!
    Word spreads so quickly. I do appreciate that reply Helen, though not addressed directly to me.
    WW
     
  51. In all of this, ask yourself the following: "If I bought a lens and it worked as expected, would I post about this in a forum? And if I did - not likely, right? - how passionate would I be in my posting and how often would I return to post about it again?"
    Then ask the following additional question: "If I had a problem with my new lens, for whatever reason, would I be more or less likely to post about it? And if I did, how might the tone and intensity of my post compare with what I might say about yet another lens that worked just like I expected?"
    The point is obvious. People are for more likely to post about a problem than about something working as expected. And people are far more likely to read and comment on the former. Almost no one among the vast majority whose new lens does precisely what they expected it to do will take the time to post: "Wow! I just have to share this astonishing news! I just got my new lens and it works exactly the way it is supposed to work! I really thank the engineers and workers at the Canon plant for doing a super competent job of producing gear that does no more or less than what I expected it to do! Let's have a 6 page forum discussion about achieving the expected! I'm so excited!"
    And if you saw such a post, tell me, would you jump into the dynamic follow-up conversation about this? Or would you either bypass it or think, "Boy, that was a strange post."
    On the other hand, can you imagine a person who thought that investing in a $2000 new super lens was going to bring them "vastly improved color and contrast," "stunningly sharp images," "unbelievably low levels of distortion," "AF speed that blows away the competition," "professional results," "unrivaled sharpness," and all the rest... and whose first photographs turned out to not meet that standard? Now this is a person who will post... early and often and with great passion! "I can't believe the low level standard the company X has for their equipment! I just spent $2000 on what is supposed to be the newest and best lens and the sharpness is no better than that of my kit lens! What a rip-off! Save your money!"
    Even more, think of the person who does get one of those rare defective lenses - and, yes, there will be some defective products among the output of any industrial process: food, cars, televisions, cookware, cameras, computers, you name it. This person, who may take this as a personal affront, will often post repetitively and passionately and often angrily. Such people make up a large percentage of those who write on forums about such equipment. (In order to avoid bringing those topics in to this discussion, I won't name them, but I think that many of us can recite the short list of "problems" that certain posters obsess about...)
    These experiences also existed before the era of the web, but the ability of the individual to spread their discontent was much more limited. In the case of a serious and real problem, the reports would get picked up by photography magazines. Camera store employees would tend to warn you away. A friend or a camera club member might say something. But the level of background noise was much lower, and microscopic and inconsequential issues rarely got blown up to the same proportions that we see on the web.
    Finally, how many of us have been certain that we were the victims of some unfair or uncaring vendor or business? And how many of us have complained not only to the business, but also in more public forums. And how many, like I have done, have at least once had to eat our words, sometimes in a very public way? Either we finally go beyond complaining to no one in particular and finally contact the company... and someone sympathizes and fixes the problem. Or, worse, as happened to me a couple years ago, we publicly and passionately impugn the integrity of the company and its employees for their incompetence and unfair and uncaring attitudes... only to eventually see that they were right and we were wrong. I'm glad to say that I learned this lesson in a very powerful way not that long ago, when I felt obligated to write a very public apology for something unwarranted that I said about a business.
    And, yes, there are some real problems - but not some scandal of generally poor quality control or companies out to rip us off. When the real problems occur, deal with them as what they usually are - individual isolated cases that need to be dealt with.
    Dan
     
  52. It's amazing to me that the sharpness and focus measurably improves at the 400 end of my 100-400 when on a tripod, with mirror lock-up, IS off (eliminates IS spin up lag and is not recommended on this lens), the lens is properly focused and apparent shallow depth of field is accounted for. I have had quality problems with third party lenses but never with Canon EOS and EF in over twenty years of use. I have broken a body or two on my own, however. I have dealt very successfully with B&H and Adorama and, in fact, have bought refurbished Canon equipment through them as I knew it went back to the canon factory. . I have never complained when one of my third party lenses failed. I have simply reverted to OEM equipment exclusively rather than use third party equipment.
     
  53. I have made purchase from both B+H and Adorama. My shopping experience has NEVER been less then perfect. They will always be my go to outlets for buying photography equipment.
    I have also made purchase from Amazon, which at times didn’t go very well.
    I wish to thank both B+H and Adorama for coming to the forum to answer a very simple question.
     
  54. PS I should have noted a 'Thank You' to Henry for alerting me to this thread.
    Helen
     
  55. Ilkka,

    I find what you say about autofocus performance varying with distance interesting, as I have noticed that especially
    with short focal lenses. Ifind it a bit more difficult to get dead sharp focus on distant objects with my 35 and 24, than
    say with my 135 (all canon), the latter being amazingly accurate in all circumstances. Maybe someone with better
    technical background than me could explain this.
     
  56. but consider this. Reports of bad lenses are not rare.​
    In comparison to the thousands and thousands of lenses out there, Alan?
    Yeah, they're rare - incredibly rare.
     
  57. I have had a lot of problems with lenses and have sent back something like 5 to date. If I recall correctly 3 of them where way blurry in one corner of the image and sharp in the others. The other two lenses did not auto-focus well, missing focus by a huge amount 80% of the time.
    I don't buy lenses over the internet any more, I want to see how they will work on my camera before buying.
     
  58. G Dan Mitchell: Or, worse, as happened to me a couple years ago, we publicly and passionately impugn the integrity of the company and its employees for their incompetence and unfair and uncaring attitudes... only to eventually see that they were right and we were wrong.​
    I have an incredible amount of respect for this post. This is the height of online courtesy and responsibility and I hold your integrity in high regard. I say this without knowing (or caring) who the retailer in question was.
    Helen Oster: I should have noted a 'Thank You' to Henry for alerting me to this thread.​
    Happy to help when I can.
    Henry Posner
    B&H Photo-Video
     
  59. Henry, thank you very much for your follow-up comment regarding my own painfully-learned lesson. To be clear, the situation I refer to (in which I wrongly and ignorantly blamed a business for what turned out to be my own blunder) had nothing to do with photography or a photography business at all and most certainly nothing to do with B&H. My experiences with B&H have been uniformly excellent.
    Take care,
    Dan
     
  60. I had 3 L series lenses. One of them back-focused quite a bit, the other two were spot on. Doesn't matter how many lenses Canon sells, in my case it was 66% good and 33% bad. I had to live with back-focusing till I got a body that was capable of micro-adjustment, which fixed the issues.
    It is for the companies to decide on what the focusing tolerances are and what percentage of finished good fail the tolerances. After all, if more lenses fail the QC test, it affect's the company's bottomline.
    Somebody mentioned - Photography is about photographs, not about equipment.
    I agree partly. Equipment does play a major role. Try focusing on an object with narrow depth-of-field with front/back focusing issues. In short to mask equipment failure, one would have to increase the depth of field.
    Do you think its acceptable to tell a soldier that the gun will fire only 99% of time.
    Quality Control comes with a price and I do expect L series lenses to have near 0 defects, considering the price paid for the lens.
     
  61. Modern lenses are more complex both mechanically and optically than their counterparts from 30+ years ago, e.g. auto-focus using internal focus units instead of manual unit focus via a simple helicoid. Autofocus and it's interaction with the camera adds a large new set of potential issues and is probably accounts for the majority of the complaints. Lenses that did not exist then are common now (e.g. 24-70mm f2.8 zooms) so internal zoom mechanics are more complex to get over 3x zooms.
    We also demand more from our lenses: not a lot of people routinely printed larger than 12x from 35mm. And if they did, they typically planned ahead, used a fixed focal length lens, a tripod, carefully focussed and stopped down. Today, just viewing an image full screen on my monitor is 22x the size of a Canon 1.6x crop sensor and it's trivial to inspect images at any magnification I desire. Many users now expect that a hand-held auto-focus image shot nearly wide open with a zoom lens will easily go 20" wide and more.
    So I'd say QC is probably better because it has to be.
     
  62. Jay. I worked in aviation R&D for several years specifically with landing and navigating aircraft in low visibilty and using GPS for this task. Even the best of equipment fails. It becomes inaccurate and sometimes unreliable or has production faults. This is true with all things we make. There is no such thing a hundred per cent. Our standards were five nines or .99999 reliability for landing systems. How we made our systems safe was to guarantee that if there were a failure that the aircrew was informed in less than seven seconds, that the system failed passively before bad guidance had an effect, and that system failure was held against very high accuracy tolerances so as not to stray from the desired course and elevation at critical stages of flight. . Nothing that we produce is a hundred per cent. Over the last twenty years there has been plenty of evidence that Canon and Nikon stand behind their products if you send the defective product in while under warranty. I am also a former military and civilian pilot. I have seen plenty of things go wrong with aircraft over a forty year aviation career. What we make fails. What we use fails prematurely at times in operation. What we do in aviation given very high stakes is to make sure that failure is passive, gives quick warning and provides a safe alternative in recovering from a failure. Having investigated aircraft accidents I know that on occasion that despite our best effort there are sometimes terrible consequences. Having been an Air Force test pilot and Quality Control officer I know that things break and sometimes don't pass field inspection from the factory. Although anecdotes don't make good science in my own experience I have had a 70-200 2,8 non IS that I have used and abused (dropped on cement, rained on, gotten muddy etc. since 1997 in a wedding business I formed after I retired from aviation. It has been used for newspaper work including lots of sports, weddings, other commercial photography and it still works as new and looks new without any maintence for almost fourteen years. I think that L lens is a very high class piece equipment. I regularly still shoot large swim meets with it. I think that's pretty amazing considering I have seen all sorts of electromechanical devices break for well over forty years in aviation. Having been a range officer in the Air Force I will tell you a little dirt can cause weapons jamming raise hell with weapons reliabilty . We never called our weapons guns by the way. We used that term for something else. I would like to see reliability figures for something like L lens although that data would be almost impossible to collect. I think the quality of today's photography products, considering the enormous volume of somewhere around 100 million units in 2010 put on the market is quite good thanks a lot to ever improving technology.
     
  63. To further Dick Arnold's post :

    1) in my experience, it's not just Nikon & Canon which stand behind their products, but also alledged lesser players
    such as Sigma (see my previous post);

    2) we should take Dick's opinion on reliability seriously, since he's a former USAF officer : after all, Murphy was a
    USAF mechanics in the USAF during WW2. We are talking about the author of Murphy's law here : "what may go
    wrong, will.". ;-)
     
  64. Dick,
    I don't think anyone here is disputing the fact that perfection is impossible to achieve. What is at issue is whether the defect rate is high or low.
    Let me repeat a point I made earlier. If someone receives multiple defective units, and there are many such reports, including several mentioned in this thread, it indicates with a high probability that there is a high defect rate.
    I don't think you will find Canon or Nikon or the other makers releasing their quality control statistics. I will guarantee one thing however, unlike the systems you dealt with in the air force, photographic equipment is nowhere hear five nines reliable. It might be closer to 99% on average, and probably much worse for some products.
     
  65. I don't remember having any major problems with any Canon equipment(lenses/cameras) except due to user error, or not reading the manual, consider me lucky.
    With that said, historically the Japanese have a much more stringent Quality Control work-flow in their manufacturing ware houses than other countries. I learned this about the Japanese when I took a business course in high-school ages ago.
    Unfortunately I can't say I had such luck with lighting equipment purchased from long standing companies with solid reputations in the industry based right here in the US(Paul C Buff excluded). Not wanting to name any names, I once purchased 3 monolights from one of these manufacturers and had to return all 3 of them due to malfunction !
    Another time I had to return the DC battery pack I purchased from the same company. My nature is to forgive and forget especially when the replaced item does what is supposed to. I figured it was just a glitch, so as I kept building my lighting system, I kept running into more problems.
    *Each time I had to return an item it cost me extra shipping charges and more than a few grey hairs.
    Recently I purchased 2 flash-heads(non mono). On both units, the clamp screw to attach the light to a stand were not long enough, so I had to call the manufacturer and order longer screws, this was provided for free, but it makes you think are these guys really Testing this equipment before it leaves the warehouse ?
    I also purchased 2 5" reflectors separately for these flash-heads, which were supposed to easily clip on the head by the use of spring-screws. Unfortunately, the spring-screws on one of the reflectors was missaligned and did not fit properly onto the flash-head. Frustrated I managed to stretch the holes on the head with a screw driver so that the spring-screws could fit right in. luckilly that worked, because I just could not take sending another item back to this manufacturer, it was costing me too much.
    So to answer the question do I think QC(quality control) is decreasing ? It depends on the Manufacturer, some have not adapted well to the current level of volumne, that they have to deal with now.
     
  66. ShunCheung

    ShunCheung Administrator

    I will guarantee one thing however, unlike the systems you dealt with in the air force, photographic equipment is nowhere hear five nines reliable. It might be closer to 99% on average, and probably much worse for some products.​
    We need to keep in mind that it is niether in Canon, Nikon ... Adorama, B&H as well as your interest to achieve "five nines," i.e. a 99.999% reliability, nor a reliability as low as 80%.
    To achieve 99.999% reliability, you must have very thorough testing and multiple backups. In other words, if you need one 1Ds Mark III, you may need to buy three of them, and each unit that leaves the factory may have to go through several hours of testing. That kind of reliability is very costly. Are you prepared to pay say $50K for a Canon DSLR and alwyas carry a couple of backups around? Of course not.
    On the other hand, if 20% of the new cameras require warranty service within the first year, that mean Adorama and B&H will receive a lot of returns and exchanges; that is costly to them. Canon may need a huge warranty service department with lots of technicans to perform those free repairs (at least at no additional cost to the customers); that is expensive for them.
    That is why most manufacturers stike a balance between those extremes. Typically, a 3 to 5% return/warranty repair rate in the first year is normal. Anything much higher or much lower will be costly, for different reasons.
     
  67. One of the mantras in the field of quality assurance is that the return for greater quality is higher than the cost. In other words, quality pays. I suspect there is a lot of truth to this.
    Shen Chung, if your estimated return/warranty repair rate is 3 to 5% is are dominated by defects, then quality could not be defined as "high", at least not in my view. It might be an "acceptable" quality standard by some measure (probably as defined by the company, since they control quality) but it means that the risk or receiving a defective product is not particularly low.
    For example, as a consumer I would not feel happy about spending $1000-2000 on a product with a one chance in twenty of being defective. I might be willing to accept the risk if there is no alternative, and I might be willing to buy the product if I don't know the defect rate is so high, but I would not normally feel good about knowingly spending my money on a product with a defect rate as high as 5%.
     
  68. ShunCheung

    ShunCheung Administrator

    For example, as a consumer I would not feel happy about spending $1000-2000 on a product with a one chance in twenty of being defective. I might be willing to accept the risk if there is no alternative, and I might be willing to buy the product if I don't know the defect rate is so high, but I would not normally feel good about knowingly spending my money on a product with a defect rate as high as 5%.​
    Alan, first of all, just use my numbers of ballpark guidelines, not accurate figures.
    If you are not happy with 5%, I can bring that down to 2%, but the cost doubles. As a result, say a 17-40mm/f4 lens will cost $2000 instead of $1000. It'll probably become very heavy and bulky so that it can tolerate some abuse. Are you willing to accept that? If you want to bring it down to 1%, it might become $4000.
    When I travel, I bring a minimum of 3 DSLR bodies so that I have at least 2 backups because I want effectively 100% reliability. I also bring backup lenses. But I admit that I am very demanding; most people wouldn't pay for that.
     
  69. Alan,
    Did you read the link I posted, twice? Is that not a perfectly reasonable assessment of the situation?
    Do you realise that now with the enlargement ratios we take for granted, but never print, are showing up engineering tolerances far beyond the limits we ever saw in film cameras? Isn't it strange that the vast majority, the overwhelmingly vast majority, of complaints are glib "this lens isn't sharp" comments"? If the manufacturers quality control was that bad the IS units, the aperture units, the zoom mechanisms, seals, grips and switches would all attract as much criticism, but they don't. Seriously, at 100% on screens, things as small as a grain of sand on your lens mount can show a difference in side to side sharpness in optimum shooting conditions.
    There are some duff lenses, cameras and planes (737-300's spring to mind), there always will be, but why assume that somebody who doesn't know how to set up one lens will be any more competent at setting up the next two they get? It is like saying this pilot crashed three planes, the planes are bad, I'd say it was more likely he was a bad pilot!
     
  70. William Michael

    William Michael Moderator Staff Member

    . . . Gidday Scott,
    DXO not being able to test tissue paper, was funnier than the Pilot analogy - you must be tired today?
    Best,
    WW
     
  71. "If someone receives multiple defective units, and there are many such reports, including several mentioned in this
    thread, it indicates with a high probability that there is a high defect rate."

    Not necessarily. Let's say there are, to make up a number, 20 reports. 20 reports against a total production of 100
    units would seem to suggest a "high probability" of a problem.20 reports against a total production of 20,000 units
    would not.

    You would also have to look at possible explanations of one person receiving "motile defective units." This is not
    dissonant with the possibility of a high defect rate, but neither is it the only reasonable explanation in every case. One
    possibility is that out of the tremendous number of people buying lenses, statistically there is some probability that
    someone will get more than one defective unit. Lightning can strike twice. Another possibility is that there is something
    connected to the person. Possibilities include unrealistic expectations, poor testing procedures or analysis, purchase
    from a vendor who does not take care of equipment very well or does not properly pack fir shipping, a buyer who may
    purchase used copies rather than new or purchase from a shady source, something about how the buyer handles or
    shoots the lens, and so forth.

    My points do not disprove the theory that there is a problem, but your explanation is at least as far from proving that
    there is a problem.

    Dan
     
  72. Spearhead

    Spearhead Moderator

    I would not normally feel good about knowingly spending my money on a product with a defect rate as high as 5%.​
    You have the statistical data to know this? Where do you get it?
     
  73. I will go along with Jeff. I stated in my post that it is unlikely that OEMs will never pubish reliability figures or even have complete data on reliability. I don't go along with speculation and having been in R&D I don't go along with exptrapolating obviously incomplete information. It is hard enough in aviation to collect information in a much more controlled environment. What I do know is that in succeeding years I have seen the quality of the equipment I use continously improve. I also believe that prices for gear would be much higher if there were not the economies of scale brought about by the large production volumes indicated by annual sales. Remember that most aircraft accidents, according factual information, are caused by human error. There is no way of accounting for human error by consumers in assessing their individual purchases. Lord knows after more than twenty years of Canon use I still screw up on a lot of pictures. As I said earlier, putting one of my four L lenses on a tripod, paying attention to focus and DOF, using mirror lockup and a remote realease, getting exposure right have gotten rid of a lot of fuzziness on so called bad lenses that I have had. I have had a lot of lenses and at least ten or twelve Canon bodies over the years since I bought an EOS 650 in 1998. I do not know what product defect rates are. I could not even speculate on that knowing that any failure collection, even by the manufacturers, is withheld or actually more to the point, incomplete and there is a lack of universally used standards to determine parameters to judge pass or failure on given products. I have used a lot of gear in the business I formed when I left aviation and mostly manual MF, 35 mm got me by and satisfied my customers and digital gets me by better these days albeit it is a good bit more expensive. In my humble, anecdotal, personal, and unsupported opinion I think the reliability of today's stuff is pretty damn good. Someone prove to me otherwise. Remember that nothing is perfect and although our standards are very high in aviation remember that we spend tons of money dealing with the probability of failure and mitigating that failure by heavily backing up systems with software and, hardware. We get no where near five nines on non- critical systems in aviation but when you are doing a Category III autoland with virtually no forward visibility it is essential ot meet those kind of standards. Because camera use is not that critical standards can be significantly lower but I certainly don't know where to find them. I suppose they are kept in secure locations at the various manufacturers.
    .
     
  74. Jeff,
    The 5% figure was a hypothetical figure used for sake of discussion. It was not based on any data I have. It was based on an unofficial estimate Shun Cheung had given in an earlier post. My discussion based on this figure was contingent (see the word "if") and explored the consequences if the figure were close to the real figure.
     
  75. Spearhead

    Spearhead Moderator

    The 5% figure was a hypothetical figure used for sake of discussion.​
    That's pretty useless when it comes to quality control, which is measured statistically. "Hypothetical" data is irrelevant unless it's on an exam.
     
  76. Dan, let me go through the discussion again regarding receiving multiple defective units by a single person. If the defect rate is 1%, then the probability that I will receive three consecutive defective units is one in a million. (The figure is 1% to the third power.) Therefore, if I were to receive three defective units in a row it would be reasonable to conclude that the defect rate is probably higher than 1%.
    In a case like this 20 reports of consecutive defective units out of a production run of 20,000 would likely indicate a significant quality problem
    by the way, my discussion does not constitute "proof", but rather is more a discussion of likelyhood.
    A full and rigorous calculation of probabilities for a pool of 20,000 customers is more involved than the simple calculation above for a single customer. I will try to give some thought to providing a more rigorous evaluation of the probability calculation. However, given that the odds of a single draw of three units has a one in a million chance of having three defective units (assuming a 1% defect rate), it seems unlikely that a pool of 20,000 customers will produce even a single customer with three defective units. Therefore if one or more customers draws three bad units then the defect rate is likely significantly higher than 1%.
     
  77. Here is how to analyze the problem.
    Assume a production run of 20,000 units and that all units were sold to customers. Assume a 1% defect rate.
    Out of the 20,000 units the expectation values is that 200 customers will receive defective units (1% of 20,000).
    Now, assume that all 200 customers place a claim and receive a replacement unit. Of those 200 customers, the expectation value is that 2 will receive defective units (1% of 200).
    Now, assume that both of those customers place a claim and receive a replacement unit. Out of those two replacement units the expectation value is that 0.02 people will receive defective units (1% of 2).
    Therefore, if the production run is 20,000 and if the defect rate is 1% there is only one chance in 50 that there will be three defective units distributed to anyone.
    Therefore, if even a single person receives three defective units it is good evidence that the defect rate is higher than 1%.
     
  78. Alan,
    Yet again, your assumption is that because people say their lens is not sharp that it is defective. This has been shown by high volume users to, generally, not be the case.
    Please read the link and comment on it.
     
  79. Jeff,
    You said "That's pretty useless when it comes to quality control, which is measured statistically. "Hypothetical" data is irrelevant unless it's on an exam."
    Since the companies do not release their statistical quality control information, any discussion of this topic is necessarily based either hypothetical numbers or anecdote.
    Anecdotes that report no problems (such as "I have bought one, or two, or even several units and have never had a problem") are not very useful because defective units represent a minority of the total. Therefore, most people who report no problems are only confirming that most units are not defective, but provide little or no evidence that the defect rate is very low. For example, if the defect rate were 5% (a figure chosen for sake of illustration), and if a person buys three product, then there is an 85.7% chance that a person will receive three good units. Therefore, someone who reports receiving three good units is not really adding much useful information to the discussion.
     
  80. Scott Ferris,
    Thanks for the link to the article(s). According to my reading the core of the message relates to tolerance stackup, i.e. in a combination of two parts, the tolerances can stack up either favorably or nonfavorably.
    According to my reading, another message of the article is that if the tolerances stack up unfavorably it may produce a result that is outside of an acceptable range for some customers, particularly if the tolerances are loose enough so that (in the case of focus errors) the focus error is outside of the depth of focus of a lens at large aperture. This is actually just another way of saying the quality control of the product line is not stellar. Or to put it another way, it is a reasonable expectation that if a person buys a body and a lens they should work very well with each other.
    One more comment: Jeff criticized my use of a hypothetical number for a defect rate in some of my discussions (e.g. 5%). Interestingly, the article you linked to starts right out in the second sentence presenting an estimated defect rate of 3% to 7%. My estimate (which I actually borrowed from another poster) is right smack dab in the middle of 3-7%.
     
  81. But if our expectations and demands have reached a level of un-cost effective tolerance limits and the manufacturers facilitate adjustment via micro AF adjustment haven't they done all they can? If lenses went up $1,000-2,000 each to correct that 0.0008 of an inch down to half that how many lenses would they sell? Particularly as alomost all new bodies have micro AF adjustment.
    Another thing, my comment on why it is overwhelmingly AF "issues" that are complained about. Why should AF be less reliable than IS, or the knobs and buttons etc etc?
    All companies have problems every now and again, the 1D MkIII AF did have some issues that affected some users. Some lenses have had issues with ribbon cables, interfaces and flimsy focus parts. Some, no doubt, don't focus as well as they should or don't have flat field focus due to element alignment, etc. But I don't find the cacophony of AF "issues" to be plausible, one because there is no way people are receiving two or three or more bad AF lenses in a row, and two, I don't see why the different AF systems across the range of lenses should all prove so problematic when other, equally complicated, components of the lens are hardly ever said to be defective.
     
  82. Wow. I bet the OP never thought he'd get such a response :)
    I work in the airline industry and from the millions of flight hours that modern commercial aircraft fly and the millions of people who are transferred safely from Point A to B, there are bound to be a few incidents. The rarity of these incidents means that when they do occur, they immediately come into the limelight.
    I think it's a similar thing here. So statistically rare are defects that they are immediately highlighted if and when they occur. With the ubiquity of the internet, it's easy to let one's feelings be known to all and sundry. I also strongly believe that photographers' tools are becoming so precise that they easily display any user error quickly, and this could often be misunderstood to be a system error rather than in fact being user error.
    Addressing QC, I think manufacturers have realised that today's consumer is very savvy, has much greater and/or easier access to information and through the proliferation of social media fora and email, can easily spread that information, whether good or bad, across the globe in seconds. It is in their interests then, to ensure that QC is good.
     
  83. ShunCheung

    ShunCheung Administrator

    You can check Consumer Reports. For electronics, cameras, etc., it is quite typical that somewhere around 4, 5, 6 or 7% will require repair in a year.
     
  84. david_henderson

    david_henderson www.photography001.com

    if our expectations and demands have reached a level of un-cost effective tolerance limits and the manufacturers facilitate adjustment via micro AF adjustment haven't they done all they can? If lenses went up $1,000-2,000 each to correct that 0.0008 of an inch down to half that how many lenses would they sell?​
    And this is the point I was trying to make earlier. On the one hand, is it reasonable for the purchasor of one poor example to expect everyone to pay more for their equipment so he or she might have a better probability of the perfection that most of them already get?

    Or is it the case that provided a decent warranty system is in place, many buyers would vote for a lens at $100 cheaper which had less QA time spent on it?
     
  85. Many of you are making an assumption that decreasing the defect rate will cause costs to increase. This notion is counter to modern doctrine on quality assurance, which states that increasing quality generally results in lower costs. Whether it is actually true or not that increasing quality leads to decreasing costs I cannot say out of my own expertise, but I can say that this is the general belief among quality assurance experts, and I tend to accept their word for it.


    Here is a link to a short and non-rigorous article on the topic that can serve a first introduction to this concept.


    http://ultimategrowthstrategies.com/articles/qualitypays.html
    If this notion by quality experts is correct then decreasing the defect rate in photo equipment is more likely to lower costs than to raise them.
     
  86. david_henderson

    david_henderson www.photography001.com

    I can say that this is the general belief among quality assurance experts,​
    Yes, well they would say that , wouldn't they? I mean if practicioners argued against this we could expect their expectations of gainful employment to fall somewhat? The reality is that to a point increasing QC might reduce costs and after that point it will certainly increase them. I think that logic, which reflects a probability that efforts to reduce defects incur diminishing returns, and seems to relate to the way business works across the board, is a more powerful argument than the necessarily biased opinions of some practicioners. Indeed if the point you repeat is true it would mean that there would be zero defects everywhere because it would always be beneficial to increase QA to reach that point. I mean put like that it's a bit unlikely, isn't it?
    Where there could, conceivably be a debate is in who pays for QC? In theory, its possible that firms could volunteer to absorb increased QC cost in the belief that in the long term it would help build a competitive advantage and volume will grow. Sadly with the pressure on short term results mowadays you can be pretty sure that today, customers pay the net cost of reducing defects. Further you can be pretty sure that if it 's likely that it will take more than a year or so to recover any capital cost of a QC venture, then that initiative just won't happen in many companies.
     
  87. Alan - first of all your own calculations of three defective lenses being 1 in a million is not sound. That calculation assumes that all variables are truly independent - in reality if the same client buys from the same shop and that shop is not careful in handling then that is a covariate which will drastically reduce the chance of getting 3 lenses that the customer perceives as defective. And there are other reasons that may result in getting 3 'defective' lenses that could reduce the odds further. Another, as has been said by more than one poster, is unrealistic expectations. So the odds of getting 3 'defective' lenses starts to plummet further below the 1 in a million you propose.
    Secondly:
    According to my reading, another message of the article is that if the tolerances stack up unfavorably it may produce a result that is outside of an acceptable range for some customers, particularly if the tolerances are loose enough so that (in the case of focus errors) the focus error is outside of the depth of focus of a lens at large aperture. This is actually just another way of saying the quality control of the product line is not stellar. Or to put it another way, it is a reasonable expectation that if a person buys a body and a lens they should work very well with each other.​
    How tight do the tolerances need to ensure that any lens manufactured at any time will work perfectly with any body manufactured at any time? Manufacturing costs for lenses and bodies may not be affected by increased tolerances, but rejection rates will rocket and that will increase prices.
    I suspect you have never worked in a production environment - or one allied to sales and price analysis.
     
  88. "A full and rigorous calculation of probabilities for a pool of 20,000 customers is more involved than the simple calculation above for a single customer. I will try to give some thought to providing a more rigorous evaluation of the probability calculation. "
    "Full and rigorous calculation" based on grossly incomplete data and unverified assumptions mostly derived from hearsay on web forums is worth approximately nothing.
     
  89. Spearhead

    Spearhead Moderator

    first of all your own calculations of three defective lenses being 1 in a million is not sound​
    Exactly. Quality for a product is a statistical measure of the units that ship out the door. Basing anything on what consumers buy and say on the web has no statistical validity at all.
     
  90. Just to be clear I was an executive in the FAA when I worked on civil GPS research and procurement. I am a strong believer in drawing statistical conclusions by using a factual base and thorooughly and clearly identifying assumptions, and accounting for and defending variables affecting the process before advancing some incomplete theory. I worked earlier in the Air Force as the head of QC and flight test in one of the largests wings in SEA. I have learned to live with realistic expectations of what is technically possible not what I would like the results to be. The 737-300 is a perfectly good airplane. It was the inspections in the area of failure that were not adequate. Airplanes go through a structural cycle with every landing and take off. They continually need to be inspected and have discrepancies fixed as a result of all this flexing. As I understand it that was an area that was not looked at under the aging aircraft program instituted by the FAA after the Aloha Airlines decapitation in 1988. I have no doubt it is now. There are a hell of al lot of rivets on the dorsal fuselage that have to be looked at. Usually the way you improve QC is by analyzing previous failures and correcting process which is now going on with the 300. Initial QC is only the beginning and does not foresee ongoing failures after production. I am surprised that my fourteen year old 70-200 2.8 is still taking usable pictures with no maintenance or calibration over this whole period. Engineering life is full of unforecast surprises and unintended consequences that only come out after use. It is my hope that Canon and other mfgrs do good failure analyses. I know on my own 5d mirrors started falling out after a few years of use. Canon has done free fixes to strengthen that attachment. That does not mean it was a bad camera. Perhaps there are a number of lenses that are hitting the field with marginal focusing tolerances. No manufacturing process is perfect. More QC does not improve manufactuing process. Improving inherent production process does after it is discovered by QC. Generally QC uses sampling and can miss single anomolies. The least a manufacturer can do is correct those lenses and correct the process that allowed the defect. It appears that at least some of this has been done. Things go wrong, things break, things get damaged in shipping, and a significant percentage of things get misused. Neither airplanes nor cameras are perfect products when they come off the line. The process of revision started when my 5D when Canon got reports of failure. This is a never ending cycle. Life is full of oh sh----s, particularly when you have to do this for a living.
     
  91. Jeff,
    You said "Quality for a product is a statistical measure of the units that ship out the door."
    That may be how some persons (including both natural persons and companies under the definition of "persons") define quality, but basing quality solely on what goes out of the factory is not best practice. A better practice is to base quality on the end user experience. I could relate an anecdote. Part of my work experience has included working in a company that buys expensive scientific instruments. Out of several million dollars worth of purchase on items totally over twenty units we had to send more than one unit back to the manufacturer due to defects. At one point the manufacturer blamed the shipping company, and that hypothesis was reasonable in some cases. However, the bottom line is that some of the instruments didn't work right, and when defective units came in the door it made for an unhappy customer, regardless of whether the problem was on the factory floor or part of the distribution channel. It was up to the vendor to improve their processes, including the distribution channel.
    Furthermore, there is a closely related concept known as "fit for purpose", which basically means that a product or process or service should work as it is intended. If there is a high rate of customer dissatisfaction then there is a good chance that the product the customer is receiving is not quite fit for purpose.
    As to your comment that "Basing anything on what consumers buy and say on the web has no statistical validity at all", in my opinion that is an extreme view. Customer feedback can be a useful indicator. Let me give another anecdote. I recently attended lectures by a doctor who specializes in coagulation disorders. He is in charge of the coag lab at a large health care organization. He has said on numerous occasions that some of his clients are very astute at picking up problems that had somehow escaped the notice of the performing lab.
     
  92. Dick Arnold, you said "More QC does not improve manufactuing process. Improving inherent production process does after it is discovered by QC."
    I think what we have here is an underlying assumption that QC is the same thing as quality. (If this is not your opinion it is something that could be misread into your comments, and regardless it seems to be the opinion of some who post here.) QC is just one component of a quality assurance program, also known as QA. (This concept is consistent with some of your comments, e.g. feedback into the system to improve manufacturing.)
    There also seems to be a mistaken notion some (maybe by you, maybe not) that QC means testing of final units. That may be part of QC, but in the best programs it is only part of the process. In the better companies QC is performed at several stages in the process, not just final inspection, and it is far better to identify a problem early than late because it reduces manufacturing costs and reduces the chances of a bad product winding up in the hands of a customer.
     
  93. ShunCheung

    ShunCheung Administrator

    Many of you are making an assumption that decreasing the defect rate will cause costs to increase. This notion is counter to modern doctrine on quality assurance, which states that increasing quality generally results in lower costs.​
    Alan, that is not an assumption; it is a fact.
    Increasing quality will lower the cost, up to a certain point. After that point, to further increase quality (and reliability) to squeeze out the last few percent, the cost will go dramatically higher.
    I used to work for AT&T who needs very high reliability on their critical computer systems. The system I worked on required just one computer for the whole US, but AT&T had 4 dedicated computers for it. They had 2 side by side in Atlanta so that they always had a backup, and then they had another identical set up in Denver. If somehow the entire Atlanta site is destroyed, they could immediately switch to the Denver site. Additionally, we also had an identical set up in the labs in New Jersey that could also serve as a secondary back up. And of course there was also a lot of effort to synchronize the data between Atlanta and Denver.
    Moreover, some of those AT&T sites were built during the Cold War where nuclear attack was a real concern. Therefore, some of those computer centers were underground and bomb proof to make sure that the communication system wouldn't stop after an attack.
    When something is that critical, you try to provide the highest reliability regardless of cost. That is simply not practical for cameras as popular products.
     
  94. Shun, to paraphrase what you said, you said that after a cetain point in the quest for quality prices must increase if quality is to increase. Following up on this concept you said that it is not practical for the camera companies to raise quality without raising prices. This is really a presumption on your part. I am pretty sure that you do not know what the QA progarm is at the camera companies, or how far along they are in pursuing their quality goals, or what their return on investment is for the next increment of quality.
    However, it is also valid to point out that the best companies have programs of continuous quality improvement, and that whatever their current level of quality is it will likely get better in the future. In addition, the best companies have programs of continuous process improvement. This is closely related to but also partly distinct from continuous quality improvement, but both go hand in hand to improve qualilty and lower cost.
    Also, I thnk that your experience with redundancy and nuke-proofing in a large one-of-a-kind dedicated computer system is useful for the sake of discussion, but probably not typical of consumer products such as cameras, and the measures taken to assure quality are rather different from those used for consumer products (redundancy of units for example, and nuke proofing for another example).
    My example of expensive scientific instruments that I posted earlier is probably not fully typical of consumer products either, but it does have some similarity to consumer products in the sense that thousands of those instruments have been manufactured.
     
  95. Dick, you mentioned that you had mirrors falling out of your 5d after some years of use. From your use of the plural I assume that it happened on more than one occasion. If so, you have clearly identified an example of a consumer identified qualty problem from your own experience, in this case a flaw or weakness in the design of the product.
    This sort of quality issue is not likely to be picked up in any kind of inspection program during manufacture, but instead is going to turn up in field use of the product. However, it is still a quality issue, and I thank you for raising this example because it clearly points out that quality is not just a matter of QC at that factory and that receiving and acting on feedback from the field is important in a good quality program.
    The falling mirrors in your 5d also provides a good example in which there is little question that the fault is in the product, not in the user.
     
  96. I used to work in professional audio, selling things like multi-thousand-dollar Neumann studio microphones (the company was like an Adorama for, among other customers, the radio broadcast engineering folks). We had a period during which an unusually high number of expensive mics were getting to customers with decentered condensers and other bits of ugliness that involed return trips (through us, to Europe) for replacement and/or service. And even some of those units ended up the same way. This all coincided with a rash of other spendy items from Studer, Denon, AKG, and others exhibiting similar trouble ... and of course we all thought in terms of, "wow, quality control has really gone downhill."

    Turns out we had two warehouse guys who had developed the habit of throwing smaller product boxes, football-style, halfway down the warehouse as they stocked shelves. They also threw boxes off the dock to the UPS and FedEx drivers (one of whom finally made it clear to one of our managers what he was seeing.

    This was back in the day when there weren't widely-used discussion forums - but I can only imagine the mis-informed echo-chamber effect that would have launched a thousand wrong-headed (factually incorrect) threads about what was going on with those damn European manufacturers, blah blah blah. The "user experience" that Alan mentions is impacted by many more variables than the QA folks and policies at the factory.
     
  97. Matt just pointed out that the user experience is impacted by more variables than just the QA folks at the factory, and provided a real-live example from his own experience. This just amplifies the fact that quality is a global issue, not just a manufacturing issue, and that it makes no difference to the user whether the problem was bad manufacturing or bad distribution. A good QA system will try to encompass all quality issues to the extent possible.
     
  98. Ted, I think there's a big difference between quality control and the fact that there might be an incredibly tiny difference in the way one lens is, for example calibrated, compared to another, which would result in different results on the same camera. Same thing goes for cameras.
    I don't know if anyone has posted this yet, but it's two interesting articles about the topic of soft lenses by the guy who runs lensrental.com.
    http://www.canonrumors.com/tech-articles/this-lens-is-soft-and-other-myths/
     
  99. Matt Laur [​IMG][​IMG][​IMG], Apr 13, 2011; 01:42 p.m.
    "...Turns out we had two warehouse guys who had developed the habit of throwing smaller product boxes, football-style, halfway down the warehouse as they stocked shelves...."

    Sorry, Mr. Laur. You are a valued, contributing Photonet member but...
    The problem was much greater than the two dockworker clowns. Where was their supervisor, the supervisor's supervisor, the supervisor's supervisor's supervisor,etc? Where was the product company rep and/or the distributer's rep? It is all part of quality control. Quality control starts with the production of raw materials and ends up with the careful placement of the product in the customer's hands. The customers were right to complain to their peers. The manufacture and/or the distributor and/or the sales outfit (y'all) all share any blame directed to any one of them by the angry customers. Your company was part of the problem, not part of the solution.
    No excuses!
    A. T. Burke
     
  100. Alan. Aviation Quality Assurance (if you want to call it that that's fine with me) is holistic. It starts with manufacture and production and the subsequent monitoring and following products both by manufacturer and the FAA throughout the commercial life cycle of a commercial aircraft. When follow up issues arise on products like engines the FAA may issue an airworthiness directive for changes to that engine. Recent case in point were engine oil leaks on the new four engine airbus aircraft. I was engaged in the research, development, and fielding of the Wide Area Augmentation System which now enables aircraft to land solely on GPS in periods of low ceiling and visibility This whole system was designed with redundancy, and software where defective processes cannot back into the system. It is constantly monitored in use. Quality Control was used in the manufacture of the system. Avionics were designed against industry standards established by RTCA. The whole system of automatic fail soft shutdown was designed with continued precise quality monitoring. This is on-going QA. What I meant to say is that in aviation when I was responsible for QA in the AF (we called it QC then) I had a team of about 25 senior NCOs that monitored maintenance quality for about thirty aircraft flying in combat. We knew we could not inspect quality into the maintenance process. Quality had to be inherent in the work done on the flight line, in the phase maintenance docks, in the supply process, and in the central control of maintenance activities. Our job was to do enough inspection to ferret out weaknesses in the system and then move to management to correct those deficiencies.. The real mission of QC is to improve the process iMO it is not to improve the product by catching and patching production errors. That is the job of testing. We did this by doing in commission spot checks on combat ready aircraft to improve the quality of individual crew cheif work assigned to each aircraft, checking on the work of phase docks to assure required inspection and maintenance was carried out, and I test flew every third phase inspection to operationally check the work etc. Good product, IMO, does not reflect good QC it reflects good work. The holistic QA job is to assure that that good work is being done against a management supplied set of work and product standards. I frankly do not know how well this is being applied at camera manufacturers. I am sure they have systems to monitor their products, failure rates, return rates and so on. As someone said, like everything else there is a curve where productivity rises with good QA to a point where productivity begins to decline with too much effort to "excessively inspect quality into the product". They certainly have a bottom line interest to maintain an acceptable (to them) quality standard. A former VP for Maintenance at United AL was a good friend of mine. We were the best of friends at dinner and the worst of enemies during the day when our interests opposed each other. He used to constantly remind me that the "best is the enemy of the good". To this day I believe him. This crap is a lot messier and much more complex than I have portrayed. The only thing I know a lot about is my own experience so I use it.
     
  101. James Meketa , Apr 12, 2011; 12:41 a.m.
    Every time I return a defective lens to B&H or Adorama, I know that it will be repackaged and resold as new to someone who doesn't know any better.
    The silence is deafening, James.
    Henry Posner
    B&H Photo-Video
     
  102. ShunCheung

    ShunCheung Administrator

    Alan, it is not a difficut concepts in manufacturing that if your quality is too low, you'll get a lot of returns, warranty repairs and unhappy customers so that eventually your company will fail. On the other hand, if your quality requirement is too high so that it is too expensive to achieve it and your products will not sell.
    The Canon 1D and 1DS series cameras and the Nikon D3 family are all highly reliable, but they are all $5000 to $8000 each and they are big, heavy, and rugged. It would make no sense to make every camera as reliable as those. A lot of consumer prefer $500 DSLRs that are small and light, convenient to carry around; they can accept a 5% failure rate. A company that only produces very high-quality products will become a niche that is hard to survive.
     
  103. William Michael

    William Michael Moderator Staff Member

    The silence is deafening, James.​
    The repercussions might be louder.
    I doubt that it is only you and I who await a response from Mr Meketa, to back up the statements he previously made.
    WW
     
  104. Mr. Meketa....
    1. How could you possibly have come across actual knowledge of what the internal return policies are for B&H and Adorama?
    2. I've been in retail. In the retail business, where individual items are sold on a person to person basis, those items that the retailer accepts as a return for any reason, are in fact further returned to whomever supplied the retailer. One good reason for this is product liability. Why would I want to take a chance that a customer that I thought was all wet, might actually in fact be right, and I end up being sued for the harm done from the second sale of a defective product? I'd also lose my insurance, along with my reputation.
    Another good reason to return the product is that I insisted on having that right with every supplier contract that I signed. I had a unilateral right to return any product that had received a customer complaint. For major products, where for instance, I was an or the authorized repair station, my crew might repair or rebuild the product to meet the specifications. That product was then sold as a used product at a used price, and my supplier contract credited my account with both the repair cost and the difference in sales price.
    Additionally, most all suppliers not only credited the full value of the return, but added a fair fee for our handling and trouble.
    3. I've corresponded with Mr. Posner many times, starting back before photonet. He's a bright and capable guy with an appropriately responsible position at B&H. He's been there a long time, so he's obviously following corporate policy when he supervises or authorizes returns and adjustments. Anyone who has been a photonet member for a number of years has seen where he's stepped in to make sure his company treats a customer right, but yet has not been a sucker for scam artists. It wouldn't make sense to me that a company that would hire and retain Mr. Posner and back his customer service decisions year after year would do an underhanded chiselly thing like putting a defective product back on the shelf. Heads of companies tend to hire people like themselves. Because of Mr. Posner's tenure, I presume that B&H ownership and top executives (which may now include Mr. Posner) think at least somewhat like Mr. Posner.
    4. Ms. Oster is an unknown quantity to me except for her responses to my fellow photonet members. Her positive answers and followup speak well for her employer, Adorama.
    5. Often criticized for it, back when in business, I was a vocal supporter that good morals was good business. If you look at an old photo magazine from 10-15 years ago, you see many, many multi-page advertisers. Most of them are now out of business, or now operating out of a mini-warehouse in the wrong end of town, and can barely finance a classified ad in the freebie local rags. B&H and Adorama remain major players in both the New York and world-wide markets.
    Could there possibly be more than a coincidence between their survival and their business ethics?
    A. T. Burke
     
  105. A.T.: You missed my point, or I didn't make it well enough. That was not an excuse, it was an illustration of a problem... but a problem well downstream from the factory. The point is that the problem (with the expensive, knocked-around studio microphones) wasn't that the factory shipped poorly made or roughly handled products. The problem was that people working for a distributor/retailer (my company) were beating up the products. This was discovered (late) and eventually stopped.

    The manufacturer clearly came to understand that their products were being roughly handled somewhere along the line between the factory door, the several trucks/docks that handled them, the air cargo system that then processed them, customs, the several more trucks and intermmediate warehousing that again touched them, the distribution and retail chain (more hands, more trucks) that then handled them, and the more packers, drivers and dock workers involved in the final shipment getting them to the end-user engineers who ordered the products. It's very difficult to isolate which of those dozens of third-party hands didn't treat a delicate product with appropriate care.

    QA at the factory (which is the kind of QA that a typical lens complaint forum post seems to have in mind) doesn't have much immediate influence over the long tail end of that distribution chain. I cited that anecdote to point out not that the warehouse management at the time was doing its job (it wasn't), but that the manufacturer still has to mop up after a lot of things that it can only indirectly influence (say, through termination of dealer agreements when a retailer refuses to clean up its act). Having spent many years in businesses selling things that people like to talk about in web forums, I'm comfortable saying that most consumers think "bad lens, factory touched it last," and truly don't understand how far away in miles, hours, trucks, packers, and modestly paid workers that factory actually is.
     
  106. At one point the manufacturer blamed the shipping company, and that hypothesis was reasonable in some cases. However, the bottom line is that some of the instruments didn't work right, and when defective units came in the door it made for an unhappy customer, regardless of whether the problem was on the factory floor or part of the distribution channel. It was up to the vendor to improve their processes, including the distribution channel.​
    Alan - does that not make CaNikon's acceptance of errors beyond their control to be even more admirable and evidence of great customer care, even of rissues beyond their QC (which, after all, was the ponit behind the OP)?
    Customer feedback can be a useful indicator.​
    Agreed, and what sounds and acceptable theoretical limit of quality can turn out to be a pin in the neck in the real world, and the manufacturer will adapt accordingly or go out of business.
    But following yor analogy, scientific instrumentation either measures known standards or it does not ( I spent years in a research labs so ahve seen this myself) - it is largely an objective assessment. But this thread has two important differences: firstly we are talking here about a subjective definition of whether a lens meets the end user's visual expectations. Secondly, a scentific instrument as you describe is in many cases designed to work as a standalone unit whereas a lens has to work with any one of millions of other units (namely, a body of any age and any model).
    Add to that, 'acceptably sharp' to some are 'total crap' to others, and others have totally unrealistic expectations when viewing images at 100% on a computer screen (equivalent to a 6-foot print from a 35mm sensor). But you seem to be ignoring that inconvenient fact.
     
  107. @A. T. Burke
    Thank you for your kind words, which are very much appreciated.
     
  108. I sent my 17-40L lens to the Canon Service Center as i was not getting good results with it. They wrote back saying they had to do some adjustments and when i got the lens back it was much better with IQ noticeably improved. I havnt had to do this for any of my other canon lenses.
     
  109. Mr. Laur.

    Thank you for your well-thought-out response. You did indeed illustrate the problem well. In your case, odds are that your company's personnel did the majority of the damage to what would have otherwise been a quality product. However the question still remains: why was the problem not caught and corrected due to the efforts of supervision, management, and ownership in your company? Your company only found out because some third party told you. I grant that the reaction was apparently appropriate and that the damaging behavior was stopped.

    Additionally, you say other quality products you sold made by top brand companies were also exhibiting similar troubles at the same time. Why would a competent group of supervisors and managers presume it must be the fault of all these top name companies and none of their own? If the owner of the company was called in for a parent/principal meeting regarding his child, would his answer be "not my little Johnny, he'd never do that."?

    I'm not trying to infer you are in a chain of responsibility, either directly or indirectly. Maybe you were the bookkeeper. But to your employer, I still say, "No excuses."

    As to the manufacturers, depending upon the jurisdiction, they probably had little or no product or mercantile liability. They might well have even been excluded by the judge if named as one of the parties in a lawsuit by a dissatisfied customer. However, these top companies are saying, directly or indirectly, to their customers, "Buy our product so you can own, use, and benefit from our high quality gear." That wasn't happening.

    Manufacturers sales reps are the people who are supposed to look into problems like product quality control. Some of them and their supervisors and managers have lost sight of that in their quest to push sales numbers. Of course, when the manufacturer is blamed, whether directly responsible or not, it does not help produce sales increases. Few sales managers get it. No excuses.

    If a company wants to maintain a good quality reputation, they've got to pick their employees, their suppliers, their shippers, their distributors, and their final sales outlets carefully. They must also monitor the chain to see that the customer gets the quality product in their hands, or they're going to lose sales. In tough times, they could lose the company. Yes, it might be a lot of work to trace quality control issues through all the various hands the product passes, but their reputation depends on it. If they don't do it, they lose customers, no excuses.

    In addition, you're probably right that the two warehouse clowns were responsible for the multiple line quality control problems. But..in fact, neither one of us knows whether they were the sole contributor, or just the last contributor and the one that got caught. If I were a lawyer, I'd hate to take a case where I had to prove the condenser was centered when the warehouse clown kicked the box.

    My point is simple: the customer's responsibility is to pay with good funds and to not blame his own mishandling of the product on the supplier. Everything that happens to the product before it safely reaches the customer's hands is the responsibility of all the people in the design, production, and distribution chain. I hold them all responsible, no excuses.

    The engineers you referred to might have blamed the manufacturer. Other engineers might blame the shippers. Other engineers might blame the custom houses. Other engineers might blame the outfit that actually sold them the product. In different circumstances, one or more of the different groups of engineers will be right. But that's immaterial. It isn't their job to figure out who's to blame. All they have to do is find a new source of a similar product. It's the responsiblity of and to the financial benefit of the manufacturer to see that a product of expected and promised quality ends up in the engineer's hands. The manufacturer must do whatever it takes, which may even include terminating dealer agreements that would cause your former employer to no longer carry the product. But in any event, the manufacturer has no valid excuse.

    Thank you again for your thoughtful answer. This is not a personal attack against you or your opinion. It is just an old-fashioned, outdated statement concerning quality control. Actually, I expect very few entities from manufacture to final seller to make a proper effort to see the customer gets a quality product handed to him. All the more reason to value and continue to do business with those few companies that do meet quality control standards and take responsibility for the few errors that are going to happen despite their best reasonable efforts.

    A. T. Burke
     
  110. A quick note or two: Where I do my work we have a "quality assurance" (QA) department whose responsibility includes all aspects of making sure we do quality work. One part of QA is "quality control" (QC), which includes testing of units and interpretation of test results (Shewhart charts and the like, though in my business we call these charts by a different name) and corrective action is taken when testing indicates that a process is out of control. The QC process is primarily applied to dummy units that go through the whole process. In the part of the operation for which I have responsibilities we also apply certain tests to each and every unit produced, and each unit has to pass certain criteria or be rejected. This is in addition to that part of QC used to produce Shewhart charts.
    One might guess that it is very expensive to test every unit. However, in the context of my work it is not particularly expensive, probably of the order of a few cents per unit, but certainly well under a dollar per unit.
    In addition, there is a report that accompanies each unit. This report has to be double verified, which basically means it has to be checked by at least two individuals. The financial cost of doing the double verification is process is less well under a dollar per unit as well.
    There are additional measures we take as well. For example, we have monthly meetings where quality indicators for all departments in our section are reviewed and discussed.
    We also undergo external checking of our ability to do quality work, and we are also regularly inspected by outside agencies to make sure we are doing quality work. (These two things may sound the same, but they are not.) If these processes turn up persistent problems then there is the possibility that the operation may be shut down, either temporarily or permanently, and there is even the very real possibility that managers could go to jail.
    All this may sound expensive and fanatical, but it is not too expensive because the quality program is built in an made highly routine, and furthermore, the consequences (both financial and other) of a quality breakdown can be very serious and potentially affect the well being of of many people. You can believe me when I say that you should be glad we pay fanatical attention to doing quality work.
    QA is a comprehensive program which aims, among other things, to minimize the number of repeats. It includes, for example, measures to build processes that minimize the chance for human error or other sources of error that can cause QC failure. In the long run this saves money because if there is a QC failure we have to send units back through the process, which incurs additional cost. One example of a QA measure is to upgrade processes to be robot based rather than human based. There are many other aspects to it as well.
     
  111. Alan - The process you describe if pretty standard QC stuff for any premium brand company and I am extremely confident that Canon have a similar QC process to the one you describe. I think I remember that you work in an analytical lab - but even there the quality and reproducibility of the results will depend on what you are measuring, what the client wants the results for and how much they are willing to pay. If they want highly accurate results reproducible opver long periods of time then they kno they will have to pay to get them. If they want semi-quantitive results it is much cheaper. The client has a choice. The same thing goes for cameras - if I walk into a camera shop and want a lens that I am guaranteed will produce absolutely pin-sharp plane-of-focus results with any body I care to put it on, then I had better be prepared to pay for the higher manufacturing costs that will entail (including the far higher QC rejection rate). Or do I accept there will occasionally be lens-body mismatches due to the QC range that Canon define?
    I have worked at different times in labs and in manufacturing and although the process of QC can be very similar, the application in practice can vary quite a lot. In the quality meetings you mentioned, I bet there have been several where the general theme has been 'we have received two complaints, nothing to worry about so we will not act yet but we will keep and eye on it'. If that is the case then you too are setting a trigger as to when you will act - just like Canon, Nikon or whoever.
    You referred previously to the acid test of client satisfation and I think it is there in their market standing. Let us set this nightmare scenario: Canon get complaints from their clients about poor quality but to keep manufacturing costs down Canon decide to do nothing - do you really believe they would maintain their position of one of the top two companies in the world? In this sort of field, once you get a bad reputation among people who need your goods to earn a living, doubt and insecurity are absolute killers and they start to drop you like a hot potato and you slide down the rankings pretty damned quickly.
     
  112. Mike,
    Thanks for the comments. Just one comment in return. When we receive a client complaint it normally triggers an investigation, and the results of the investigation are reported at the monthly meeting. It is taken very seriously and I don't recall seeing a case where the attitude of "nothing to worry about so we will not act yet but we will keep and eye on it" prevailed. Although I have not worked at our competitors, I think their approach to client complaints would likely be taken similarly seriously.
     
  113. When we receive a client complaint it normally triggers an investigation, and the results of the investigation are reported at the monthly meeting. It is taken very seriously and I don't recall seeing a case where the attitude of "nothing to worry about so we will not act yet but we will keep and eye on it" prevailed. Although I have not worked at our competitors, I think their approach to client complaints would likely be taken similarly seriously.​
    +1!
    Henry Posner
    B&H Photo-Video
     
  114. Us too!
    Helen Oster
    Adorama Camera Customer Service Ambassador

    Helen@adorama.com
     
  115. When we receive a client complaint it normally triggers an investigation, and the results of the investigation are reported at the monthly meeting​
    You correctly review and investigate complaints and you then make a decision whether or not to take action. So my (poorly described) point was do you take remedial action every time you receive a complaint. Do you change process every time you receive a complaint - and that is the intimation behind comments about Canon's QC. .
     
  116. My earlier post (April 12) about returns being recycled was based on what I was told explicitly by the owner of a large, east coast camera store when I asked him what would happen to a brand new, but badly decentered, prime lens I was returning. I assumed from his reply that this practice was standard. I am glad to learn that it may not be.
    As far as Adorama and B&H are concerned, I have purchased large amounts of equipment from them over the past two decades. Both are fine retailers, and I have never had a single problem with either one. I can recommend both highly.
     
  117. William Michael

    William Michael Moderator Staff Member

    Thank you for answering my question.
    I think it was either an unfortunate mistake, or gross misjudgement, to assume that what any one business does, is applicable to all businesses.
    Personally, I believe that my many of my Customers view our competition’s practices and that is why they choose to do business with us. I would expect that most Customers do – that’s one main reason as to how they choose with whom they do business.
    I hope that your Customers do not look at your Competition - and then judge your business by your competitions practices or what your Competition’s CEO tells them: that would be just unfair, wouldn’t it?
    WW
     
  118. James
    I'm delighted to read that you've never experienced problems with purchases from Adorama, but your response highlights how reputable retailers can be adversely affected by others.
    The retailer you spoke to may well be one of those who would in fact return poorly performing units to the manufacturer as overstock - rather than as faulty.
    These same units could so very easily end up in our inventory with our next shipment from the manufacturer. As we don’t open and check units prior to despatching them to the customer, it would appear that we were sending pre-owned goods to our customers, intentionally!
    Helen Oster
    Adorama Camera Customer Service Ambassador

    http://twitter.com/HelenOster
     
  119. ShunCheung

    ShunCheung Administrator

    My earlier post (April 12) about returns being recycled was based on what I was told explicitly by the owner of a large, east coast camera store when I asked him what would happen to a brand new, but badly decentered, prime lens I was returning. I assumed from his reply that this practice was standard. I am glad to learn that it may not be.​
    James, could you share the name of that "large, East Coast camera store"? Apparntely that is neither Adorama nor B&H.
    If that is true, as we pointed out before, it is a very stupid way to do business. The chance is that the next customer will discover the same problem and will return that same item. Every time an item is returned due to defects, it generates more overhead that costs the store as well as Canon/Nikon. The only companies that may be benefitted would be UPS or FedEx if we are talking about mail ordering. I certainly wouldn't take my chances with such stores; therefore, I am sure a lot of members here would like to know which store that is.
    I have bought 40+ new Nikon lenses in the last 3 decades plus some more new ones I received from Nikon through photo.net for testing. So far I have had exactly 1 that required warranty repair within the first year (that was back in 1990) and none defective out of the box. Recently a friend and I both purchased refurbished Nikon 70-300mm AF-S VR lenses from B&H, and both refurbished lenses were defective. In both cases B&H paid for return shipping and promptly refunded us. I only have personal experience with those 2 samples among, but according to Henry Posner, that is uncommon for refurbished lenses.
    However, I have seen a few people here on photo.net. For any item they purchase, regardless of whether it is Canon or Nikon, camera bodies or lenses, they have to return everything once or twice, sometimes more, before they are happy. In one case one of those people posted image samples of his "defective" lens to the forum, complaining that one side was unsharp, and everybody else thought the lens was perfectly fine. Those who keep on coming up with those non-existing "problems" merely jack up the cost for all of us. Fortunately, as far as I can tell, there are very few of such folks.
     
  120. One problem with the discussion so far is the lack of actual data on defect rates for new photographic equipment. Some estimates have been given, and it has also been pointed out that the manufacturers do not publicly disclose defect rate data, but no real data has been given.
    To help put the conversation on a more quantitative footing I have searched for defect data, and so far I have found one web page giving some survey results for defect rates for lenses, though not for all photographic equipment.
    There are a number of limitations to the study, of which I won't go into in detail. The web page itself discussed study limitations in more detail. I will briefly summarize. The main limitation boils down to biased sampling. The survey probably over-represents users who received defective lenses. On the other hand the survey is limited to out-of-the-box defects and also excludes recalls. From the user point of view the latter groups of defects are little different from the defects that are covered in the survey, so in that sense the survey under-represents the defect rate.
    Here is the link to the study and a summary of the primary findings.
    http://www.lensplay.com/lenses/lens_defect_results.php
    Here are the results as of today. I have converted the data from the probability of getting a good lens to the probability of getting a defective lens:
    Canon lenses - 13495 with 1051 defects
    Defect rate 8 %
    The probability of getting 5 good Canon lenses in a row is 67 %
    Sigma lenses - 3073 with 686 defects
    Defect rate 22 %
    The probability of getting 5 good Sigma lenses in a row is 28 %
    Tamron lenses - 1527 with 229 defects
    Defect rate 15 %
    The probability of getting 5 good Tamron lenses in a row is 44 %
    Tokina lenses - 516 with 88 defects
    Defect rate 17 %
    The probability of getting 5 good Tokina lenses in a row is 39 %
    Nikon lenses - 2470 with 204 defects
    Defect rate 8 %
    The probability of getting 5 good Nikon lenses in a row is 65 %
    Pentax lenses - 1688 with 109 defects
    Defect rate 6 %
    The probability of getting 5 good Pentax lenses in a row is 72 %
    Minolta/Konica/Sony lenses - 669 with 62 defects
    Defect rate 9 %
    The probability of getting 5 good Minolta/Sony lenses in a row is 61 %
    Other lenses - 1007 with 106 defects
    Defect rate 11 %
    The probability of getting 5 good Other lenses in a row is 57 %
    The most striking thing about the study is that if the results are anywhere near representative of the true defect rates then quality assurance and quality control from the manufacturers ranges (in my opinion) from very bad to horrible, much worse in fact than the most pessimistic estimates given in the thread so far.
    If the data in the survey are not representative of the true defect rate then it would be very helpful if the manufacturers would release their data to the public.
     
  121. "if the results are anywhere near representative of the true defect rates"
    That has been my point all along, I am sure that lenses people say are defective are not, and very high volume users experiences back that up. LensRentals suggest their return rate across all lens manufacturers is 2%. And they admit that they won't stock many Sigma lenses because of unacceptable failure and quality rates.
    " But after opening some 8,000 shiny new lens boxes I can assure not all of them are. Whether its quality control at the factory or getting knocked around in shipping our experience is about 2% of new lenses need to be exchanged"
    I give a lot more credibility to professionals who rent lenses out for a living and have owned over 8,000 of them, than some random internet user.
    Furthermore, their experience is
    "if you send a lens in to factory repair with “This lens is soft” as the only description of the problem, chances are extremely high that it won’t be fixed. Trust me on this. We have 20 lenses a week go in to factory service. We’ve learned."
    Twenty lenses per week, that is more than most of us will do in our lives! Nobody has addressed my earlier comments about different kinds of failure either, if QC is so bad why don't IS units fail as often, or aperture units? It just doesn't make sense. Indeed I would say the lack of failure comments other than focus points to the truth being that QC is very good and most complainers are just wrong.
     
  122. Spearhead

    Spearhead Moderator

    If the data in the survey are not representative of the true defect rate then it would be very helpful if the manufacturers would release their data to the public.​
    There is no statistical validity to the tests. In addition, it ignores all the post-manufacturing issues. No controls, no good data, no value.
     
  123. ShunCheung

    ShunCheung Administrator

    Alan, that lensplay.com article openly admits that they did not sample random users. As soon as you see that, you might as well stop reading the rest of the article since it is not a statistically valid survey.
    I have been buying Nikon lenses since 1977. Their results totally contradict my experience over 3 decades. Just within the last 7 months, I have bought 5 new Nikon lenses; in fact I received 6 since I accidentally double ordered the TC-20e iii teleconverter. All 6 lenses are fine. I carefully compared the two teleconverter smaples (from different stores), and they were practically identical. In that period I also bought a refurbished lens that was unsharp on the long end, and I returned it.
    However, I am a bit surprised by the 2% reported by LensRental. That sounds low to me.
    Again, Consumer Reports publishes repair records for a lot of products, not only limited to cameras and lenses.
     
  124. OK guys, I have tried to interject some real results into the conversation. If you don't like those results, then how about coming up with some alternative results. I mean some real data from some source, rather than just carping and complaining about the quality of results I pointed to.
    The results from the Bob Atkins survey (might as well call it what it was, and by the way Bob, where are you... why not weigh in on the present conversation?) may have some weaknesses, which by the way were acknowledged in both the survey and in my post pointing to the survey, but at least they are real results, as opposed the the hypothetical speculations that have constituted most of the discussion in this thread.
    I take as an example the unsupported speculation that most of the returns are from users who don't know the difference between a sound unit and a defective one. (Where is the data showing that those users don't know the difference between a good lens and a bad one?) No data=speculation, regardless of whether the speculation is true or not.
    As to the comment that the Atkins data have no statistical validity, that is a rather extreme view. There certainly is some validity to the data, though the quality is not as high as in a well-controlled study. One thing is for sure, it is better quality than comments that are backed up by no data at all.
    Scott at least has posted a link that at supplies some alternative data. That data could also be subject to some criticism. For example it says "about 2% of new lenses NEED to be exchanged." [emphasis added]. The problem here is the definition of the phrase "need to be exhanged." Does that refer to some minimal standard of acceptability for the purpose renting a lens to a relatively non-discriminating user, or does it refer to a higher level of quality? We don't know. In any case, let us take the value mentioned in that link (2% defects) at face value with respect to the rental agency.
    My guess (yes, it is a guess) is that the true defect rate probably lies somewhere between the rates quoted by Scott's link (2%) and rates found in the Atkins survey (typically around 10%), which conveniently enough would put it roughly in the range that Shun suggested (which I think were roughly based on Consumer Reports).
    However, let's assume the defect rate is at the low value (2%). Let me then submit a question for discussion. Is a 2% defect rate is acceptable for a consumer product? I can only speak for myself that it is about an order of magnitude greater than an acceptable defect rate. I can tell you this much. If defect rates were that high at the place where I do much of my work we would shortly be out of business.
    Let me also point out another concept that has escaped discussion so far. It is the flip side of a concept that has been introduced, which is that most returns are allegedly made by users who do not know what they are doing. If this is indeed true, then consider also the flip side of this, which is that most users are also not likely to pay close enough attention to notice that a lens (or other product) is defective. In other words, most consumers will just use the product without further thought, and will not return a unit as defective unless it is very obviously malfunctioning. Thus, a lot of defects never generate a return, which would then bias the statistics (if one could ever find them) toward the low side.
     
  125. Alan,
    I have no affiliation with LensRentals, however I have both rented from them myself and used lenses that have been rented by others from them. I know there customer service is second to none, they are very accommodating and protective of their brand name.
    I believe their figures, their reputation depends on it, their livelihood depends on it. They are also very aware of potential lens problems and are set up to test their inventory when they get it. They have nothing to lose by understating their figures, they have no axes to grind and having read most of the articles/posts on the site they come across as straight talkers. Indeed there is no advantage to them keeping even a questionable lens, it will cost them money and goodwill from unhappy customers and it costs them nothing, effectively, to return an unsatisfactory new lens. I remember having read in another post that the figures are seriously skewed by including Sigma lenses, so I believe the out of the box failure rate of new Canon lenses could be well lower than 2%.
    Whether other failures are not observed by other owners is statistically irrelevant. LensRentals figures show that from 8,000 lenses around 2% across all manufacturers need to be returned, they are not missing bad lenses, even if others do. So we have an uninformed, untrained mass with inconsistent testing methods and opinions that might be complaining about 10% of lenses, or a lens business that consistently tests all their stock the same way with trained staff and acurate setups that reports an out of the box failure rate of 2%. I know which figures are easier to present as realistic.
    So your new question is, if the Canon contingent of that 8,000 lenses is showing a 2% out of the box failure is that an acceptable rate? Well it seems to be inline with the other manufacturers, Canon are not reported as being worse than Zeiss or Nikon. So there would seem to be an industry standard, probably one that comes from a balance between manufacturing costs and sale price.
     
  126. Spearhead

    Spearhead Moderator

    If you don't like those results, then how about coming up with some alternative results​
    Quality control needs to measured at the source. The introduction of variables after the completion of the manufacturing process, and the lack of controls after the manufacturing process, make everything measured beyond that point statistically invalid. So the only data that can really be used would come from the manufacturers, or someone who has done the quality measurement for the manufacturers. There is nothing else that is statistically valid.

    My father wrote the statistical training program materials for the American Society for Quality for many years and I was his proofreader. I learned a lot about this topic and how professionals view quality, and there is nothing in display here on this post that reflects that. Sorry...
     
  127. Jeff said that quality control needs to be measured at the source. I won't dispute that quality control needs to be measured at the source, but that ignores a some key issues.
    "Quality control" is only one part of making sure that the company is delivering quality products to customers. There is a whole zoo of vocabulary terms that attempt to get at the bigger picture: "quality assurance", "total quality management", "quality engineering", and so forth. However, the big picture is that the best companies try to make sure that quality refers not just to what leaves the factory floor, but rather what arrives in the hands of the customer.
    Furthermore, even if you restrict the vision just to what leaves the factory, "quality control" as it is often defined does not encompass the full quality program of a company. Quality control, as it is commonly defined, refers to testing programs and so similar activities that determine defects and prevent the shipping of defective units. This is an important part of quality programs, but equally important are other activities, such as engineering of products and production processes that make it possible to manufacture defect free (i.e. low defect rate) products.
    Going back to Jeff's point of view that quality need only be concerned with what leaves factory floor I can only say this: Such a point of view makes quality control entirely irrelevant to the customer. The customer does not care one whit about what the quality level is when a product leaves the factory floor. The ONLY quality issue a customer cares about is that a quality product be delivered into his hands, and there are quality programs that take this into account.
    One more comment to Jeff about your last post. You may have learned a lot by proofreading your father's manuscripts on quality control. However your statement "I learned a lot about this topic and how professionals view quality, and there is nothing in display here on this post that reflects that" is really extreme and not at all realistic. It may be that other posts do not conform to your view of what quality is, but this is more a reflection of your limited view of quality than a reflection that none of the other posts here reflect an understanding of quality.
     
  128. LensRentals figures are not percentages derived from the factory gate, they are delivered to the end user, customer, figures.
    Yet another unnecessarily argued point of semantics.
    Of more statistical interest is the discrepancy between a professionally tested high volume scenario, 2%, and the other, internet user, figures of 10%. I think that proves that many people who complain about their lenses, and return them, and post about it, do not have faulty lenses at all. By inference 80% of the complaints are not accurate.
     
  129. Let me make another point. I have mentioned this before, but I should bring it up again and elaborate on an essential subpoint ot two.
    Earlier I made the point that among quality professionals there is a prevailing view that quality pays. Others here have responded with the idea that increased quality may lead to higher sales and that quality pays in that sense. I don't dispute that higher quality may lead to higher sales.
    However, more important is the fact that a good quality program pays because it typically lowers manufacturing costs by decreasing defects and by catching defects earlier in the manufacturing process. It is better to catch a quality issue at the level of a $1 part that goes into a $100 product than to wait and test the product at the end of manufacture and have to reject the whole thing.
    Also, the tone among some posts have been that increasing quality is primarily based on more stringent testing, presumably of the final product, and that this is very expensive and leads to higher prices. However, more important is that a good quality program catches problems earlier (as just noted) and that a good quality program, when coupled with a good design and manufacturing program, will results in improvements that make it so fewer faulty units are made in the first place.
    As an anecdote let me relate that a few years ago I had a large number of discussions with a fellow we will call Mike about quality programs and the related topic of continuous process improvement. Mike had formerly worked for a company that supplied air bag components for automobiles. According to him, he and others in his company were taught by one of their clients (a very large Japanese auto maker) how to improve the quality of their operation by introducing a program of continuous process improvement. According to Mike, as they followed this program over a period of time their company's productivity (on a per employee basis) improved several fold, their costs per unit went down, and their defect rate was cut to a fraction of its previous value. This was all done without increasing staff or manufacturing floor space. This anecdote illustrates what is possible in a good quality program.
     
  130. Scott, you said "Of more statistical interest is the discrepancy between a professionally tested high volume scenario, 2%, and the other, internet user, figures of 10%. I think that proves that many people who complain about their lenses, and return them, and post about it, do not have faulty lenses at all. By inference 80% of the complaints are not accurate."
    Your inference does not follow from what is known about the problem. The difference between the two numbers may have many causes other than your assertion that 80% of the complaints are not accurate. For example, an equally valid interpretation is that that the difference between the 2% figure and the 10% figure is that it is a reflection of sampling bias in one source of data vs. the other. For example, the Atkins survey readily admits that there may be sampling bias due to over-reporting of defective units, i.e. those who received defective units are more likely to report a problem than those who do not receive defective units. It is therefore possible that ALL the defects reported in the Atkins survey represent true defects, but that those who received defective units are morel likely to respond to the survey, and that this accounts for the higher number in the Atkins survey. Other explanations are also possible.
    However, from my point of view the more striking thing about the two sets of figures is not that they are so different (2% vs. ~10%), but rather that they are roughly of the same order of magnitude. Taken together they indicate that the defect rate is rather high, measured in terms of multiples of 1% rather than in fractional % terms. Either way, whether 2% or 10%, or something in between, a defect rate measured in multiples of 1% is nothing to be proud of and also something that a consumer might well be concerned about.
     
  131. Alan,
    You say "defect rate is rather high, measured in terms of multiples of 1% rather than in fractional % terms". High compared to what? These are complex mass-produced consumer items with a wide variation in retail distribution. Can you identify a similar product in price or complexity that has a much lower rate?
    Also you say "quality pays" (w/o caveats) and later discuss a large Japanese manufacturer with a good continuous improvement program. The major lens makers are also large Japanese manufacturers who most likely have had very similar training. To assert that they have not calculated the costs vs. benefits of various QC possibilities basically assumes they are incompetent businessmen. That's not to say there is no room for improvement: continuous improvement is all about finding ways to do so. But with continuous improvement, you will typically reach plateaus where the gains become very small or the next cost increment is too high - at least until the next revolutionary idea. I'd be very surprised if Canon hasn't calculated the cost/benefits very carefully and is basically at such a plateau. We certainly cannot point to a commodity lens maker who is doing significantly better -- which suggests no one has figured out how to do it better.
     
  132. Erik Magnuson,
    You ask "compared to what?"
    Let me start by saying "compared to six sigma," the generally acknowledged quality standard among quality experts and among the companies that have adopted best practice.
     
  133. Alan, no reasonably complex (i.e. dozens of discrete parts) mass-produced product will even get close to "six-sigma." It's a slogan that stands for an approach to QC improvement. Can you think of a product that approaches this goal? Lets put it this way: any product that has a warranty repair service is unlikely to be it because return/repairs would be too small to maintain such a service.
     
  134. "Can you think of a product that approaches this goal?"
    Yes, how about air travel for a start.
     
  135. Spearhead

    Spearhead Moderator

    It may be that other posts do not conform to your view of what quality is, but this is more a reflection of your limited view of quality than a reflection that none of the other posts here reflect an understanding of quality.​

    No. None of them relate to manufacturer quality control. That's a fact. Web sites don't matter on this topic, they are not posting true quality control data. They are posting on perceived customer delivery quality. "Perceived" has nothing to do with quality control. "Delivery" has nothing to do with quality control. Maybe you should create a new term, because you are not talking about quality control. Read my father's book, you might learn a lot.
     
  136. Jeff,
    Maybe we are not communicating very well. You are at least half right when you say I am not talking about quality control.
    Here's the thing. I am not limiting my discussion just to "quality control", which has a somewhat restricted and specific meaning (usually taken to mean statistical quality control) when discussing quality, but rather I am talking about quality in its totality. "Quality control", by itself, has NO direct relevance to the end user. The ONLY thing relevant to an end user quality of the product as delivered to him/her. Quality control is simply one of many tools that can help assure that the end user receives a quality product. There is nothing wrong with quality control. It is necessary and useful, but there is something wrong with conflating "quality control" with the totality of a program to assure quality.
    Quality control is one component of a quality program, but only one component of the larger picture of delivering a quality product to the customer.
    And yes, I do know what quality control means. I have attended numerous presentations on quality control. I am presently in the process of arranging to have a professional quality control education service give a multi-day presentation on the topic where I work as part of a continuing education program. I have to sign off on quality control reports several times per month. I am board certified in a discipline in which quality control concepts and interpretation are part of the certification exam. I am well familiar with Control charts, also known as Shewhart charts or process-behavior charts. (They call the Levey-Jennings charts in my line of work.) I am also familiar with the application of statistical quality control to these charts. In my line of work this is generally done through Westgard rules.
     
  137. As the responsible executive for research, development, and acquisition of aviation navigation systems it was my job to view these processes as a whole. Our job was to deliver navigation and landing systems that met stringently developed accuracy, integrity and reliability standards. Quality Control in manufacturing was an element in the overall production, installation and use of these systems. A major issue was safety critical software QA. The software for aviation use of satellite systems is very expensive and has to be rigidly tested by highly educated software experts and subjected to extensive independent testing. Provisions for continously monitoring navigation signal quality during use through the use of independent monitors was also a part of the equation. Systems have to fail passively with user warning in less that seven seconds on some critical systems. That is part of the quality assurance equation. Systems also have to be reliable and reliability has to be monitored continously in actual use. Down time has to be held to an absolute minimum. In the event of a terrestial system failure parts to repair have to be rapidly available. In the event of space based signals. equipment aboard aircraft has to continously monitor signal performance and immediately provide for non-corrupted failure. A terrestial network of ground satellite monitors called WAAS has to provide accuracy data to aircraft using the GPS system for IFR navigation and deliver a verification signal through geostationary satellites independent of the GPS satellites used for navigation. In all ground systems failure data is collected and automatically reported so the system logistics are constantly monitored for excessive parts failure and consumption. This Alan is verifiable quality assurance from the top down. Quality control is not a religion it is part of building and maintaining an integrated system. A system has to be designed and built and integrated into product design against manufacturer and independently developed aviation standards that inherently provides for signal and equipment quality and reliability. My experience has led me away from speculating upon unverifiable estimates of failure rates as you all are doing. Camera manufacturers have most assuredly developed production specifications for their products whose tolerances can be monitiored post production. User feed back and return rates probably do not represent actual field failure rates in their totality. The most interesting facet of this thread is the variability of lens performance based upon a variety of interfaces, apparently untested, that are required in lens mounting on various camera bodies. This inferface issue introduces questions into determining whether failure rates are the fault of the lens mount or the camera mount or a compounding of errors between the two. Interface compatibility is essential with aviation systems. I have seen no data on that. Estimating and accounting for interface variables is basic to any scientific study. Alan, I respect your expertise but I am rather surprised at the speculative and somewhat constrained nature of your posts.
     
  138. Yes, how about air travel for a start.​
    Let's ignore it's a service and not product. What aspect of "quality" are you using: Arrive alive? Arrive on time? Arrive on time and satisfied with the service? You will get vastly different quality indices from these 3 different measures. What are the equivalent lens measures: Lens arrives in one piece? Lens performs within factory specifications? Lens performs within user expectations? You will get vastly different quality indices from these 3 different measures.
     
  139. Dick,
    Speculation? If you go back and read the posts I was one of the few people to have brought any real data into the discussion. The study I cited was imperfect, but it was real data and relevant to the topic because it dealt with defect rates of the product in the hands of the user. The results from that study pointed toward a defect rate of roughly 10% for lenses, with some manufacturers doing better and some worse. Even if that figure were an over estimate by an order of magnitude it still indicates a problem with quality in the industry.
    Another person cited results on defect rates as reported from a lens rental agency. That study was not subject to the same problem as the main problem in the Atkins survey, which was a potential for biased statistics due to over reporting by those who had a bad lens vs. those who did not have a bad lens, nor was it subject to the other (dubious) criticism made, which is that the people doing the reporting were not expert enough to know good from bad. That particular study pointed toward a defect rate of 2%. Again, such a high defect rate points toward a significant quality problem.
    In the absence of data supplied by the manufacturers (which they seem not inclined to provide) it is necessary to use whatever information is available, which means relying on a few imperfect studies and anecdote. The alternative is to squelch discussion altogether.
    You may have noticed that I was also one of the few to provide any mathematical analysis, which in that case was to calculate the probability of a person receiving multiple successive bad units. It is possible to question the results, primarily based on the input data (some assumed numbers), but the results nevetherless showed that reports of multiple defects should be extremely rare, and they are not rare, which again points toward a quality problem.
    Now let us take up six sigma again. A simple six sigma calculation will show that if the company is operating at a six sigma level (a rare thing, but not unheard of) the defect rate should be 3.5 DPMO (defects per million opportunities), and applying that rate to the individual components of the lens, and conservatively assuming the parts count in a lens is 300, then the defect rate should be 0.1%. All the available data, incomplete though it may be, indicate that the defect rate is an order of magnitude or more higher than this, which again indicates that the quality level is well below state-of-the-art quality programs.
     
  140. Alan. Maybe speculative is the wrong word but I would never rely on data as poorly supported as what you are using. What is the confidence level you would assign to the accuracy of your conclusions above? What I cited in my above post is some of the real effort behind safety in air travel. I agree that air travel is umcomfortable; airlines have high cancellation rates and security is a pain. But the pains that are taken to make the system safe are extensive. As to the recent failures on 737-300s it is my understanding (I am retired from the FAA) that the required extensive inspection and sampling done on all 737 aircraft since the Aloha decapitation in 1988 did not incluje that section of the aircraft where the failute happened. There are literally thousands of rivets on the dorsal portion of the aircraft . The program to inspect these aircraft is called aging aircraft. Southwest was fined heavily for not accomplishing these aging aircraft inspections on time although they were apparently not inspecting this area of the airplane under those aging aircraft rules. The nature of the failure is the propagation of cracks between individual rivet holes and the edge of the panel they secure. These are very tedious, high work hour inspections involving magnetic sampling on each rivet. Aircraft stretch and bend with each pressurization and take off and landing. These are called cycles. I think the aircraft in question had about 37000 cycles. This is quite a low number for these failures to start occuring. So this caught the industry by surprise. This is a straight QC problem that can only be solved in two ways. Inspection of thousands of rivets on a regular basis and replacement of individual rivets and panels or costly reskinning ot the 737-300 fleet. There is no allowable failure rate. I believe what I have learned working in aviation quality areas is that one cannot inspect quality into an aircraft maintenance or naviads production. I believe that QA or QC is a vauable resource in determining where quality defects lie but the way to improve product is to improve the system and inculcate pride of perfomance in those who actually do the work. QC cannot be a substitute for improved process and the eyes and ears of the work force. This also includes strengthening of production standards and tightening specifications. Look at what Ford has done in catching up with the Japanese in the quality area. It takes senior management commitment and understanding. I think senior mgt. is pretty damn interested at SW air lines right now. I have no idea how camera mfgrs. view quality issues so I would not attempt to render a judgment. SW lives and dies by their 737 fleet as it is their only aircraft. The reason I am commenting is that this is an important set of issues with me. I investigated aircraft accidents in my Air Force career and have witnessed and analyzed the results of quality failure first hand
     
  141. Let me correct a typo in my last post. The six sigma error rate is 3.4 DPMO, not 3.5 DPMO. I just can't understand how that typo got past my stringent QC system, which is based on 100% inspection (i.e. proofreading), but it did, so there you go...
    Which inadvertently brings up an important point that relates to quality, namely that systems that rely on human perfection, even at the inspection level (such as proofreading a post), are always vulnerable to error. One of the goals of a good quality system is to design processes so they are not dependent on human factors or at least to reduce the dependence on human perfection.
     
  142. Thanks Dick for your comment.
    One point you made that needs to be emphasized is that "QC cannot be a substitute for improved process and the eyes and ears of the work force." In other words, QC is important, but not enough. This principle is built into a good quality program, and I think I have made similar comments on this thread as well.
     
  143. I was considering buying the Sigma 120-400 lens and read the amateur user responses and blogs regarding this lens and they were all over the map. I was suspicious but not at the lens but rather at the letters and blogs. I bought one on the proviso that I could return it if I was not satisfied. My first attempts at testing my lens in real use conditions were not too good. I realised that on a tripod one had to shut off the image stabilising funciton for one thing. I noted that a number of the letters I had read did not mention having done this. I also found that shooting on a tripod with this very big lens with a full frame Canon resulted in tripod shake sufficient that when really critically testing this lens for what is its best quality proved to me that the lens was fine but that most of the time, it is camera/mirror shake and not the lens that is the culprit. Yes, I was using either timer or remote release. In the end, I found that I could stabilise the tripod with weight and the lens was fine and dandy. I wanted to use it portably and I solved that well also. I built for this lens a grip so that I could hand hold it with image stabilisation and can capably take long distance shots at 1/25th of a second and print such images over 3 foot print sizes consistently. What it shows is that it is often about technique and knowledge and that the inconsistency of the optics might very well have more to do with the user than the company that makes the lenses. Sure, sometimes a lens is a dud. Many of my old fiim lenses were very inferior to the present optics common today and thus I have my doubts that there is some sort of fall down on the optical side of technology today. To the contrary. It is a fall down on the user who does not really know the photographic skills that old fllm users HAD to know in order to get any decent result what with slow shutters, less than sharp lenses and available lighting conditions with slow or grainy film stock. What I suggest for those that have what they feel is a bad lens...consider a serious and properly controlled test rather than running off at the mouth about how bad this or that lens is. Rather, one might want to think..it is I and not IT that is the problem.
     
  144. Thank for your response. Many reviews explain if the lens is good or bad, not how the review was using the lens. Your example
    shows how much testing a buyer has to go though to get a decent photo with this lens. The lens makers should also provide "how" to
    get top performance out of the lens. The trial and error you went though is not what many photographers would go though with out
    just returning the lens and writing a bad review. The lens makers create ads that leads one to believe that every lens is perfect for
    every situation.
     
  145. I recently found an interesting web page. Here is a link to part of it.
    http://www.imx.nl/photo/optics/optics/page61.html
    Although it is not the particular purpose of this web page to focus on defects in photo equipment, there are nevertheless a few comments on this topic scattered through the various sections of the web page.
    Here is an example of a quote from the page:
    "The EF 2.5/50mm wide open is quite disappointing, optically and mechanically because the optical performance is reduced by large assembly errors"
    I should point out that this comment was not from some random unsophisticated user, but is given in the context of a very sophisticated and technically astute user who is reporting results of MTF measurements of a series of lenses using an optical test bench.
    This is just an anecdote of course, not a statistical analysis of a test results of a large number of nominally identical units. On the other hand, if the defect rate of the manufacturer were very low then it is very unlikely that this reviewer would have been testing a defective unit.
     
  146. In my 40+ years of buying and using Nikon, Minolta, Olympus, and now Canon cameras and lenses I have only had two lenses that caused me troubles and both were Sigma. I won't buy another Sigma unless I can first test it. I've never had a problem with another brand. Optically, a few weren't so hot but that's another matter.
     
  147. Here is an example of a quote from the page:
    "The EF 2.5/50mm wide open is quite disappointing, optically and mechanically because the optical performance is reduced by large assembly errors"
    I should point out that this comment was not from some random unsophisticated user,​
    He does not say how he could distinguish between assembly error and later shipping/handling damage. There is an effect (misalignment of elements) but there are multiple potential causes only one of which is a QC issue. He also did not tell us the date of manufacture of that lens -- which has been in production for 20 years now -- and so even if the problem was lack of QC it could be a now corrected problem. See how one example does not a trend make? Then if you read further down the page he says:
    Japanese manufacturers adopt the current approach where production aims for a certain quality level and then let quality control go for the Gaussian error distribution, neglecting the extreme 3% or 5%. It is much cheaper to replace a defective unit after a complaint than to strive for 100% quality control at the end of the assembly line.​
    3% is close to the ~2% reported by the lens rental place. Note his assertion here about costs conflicts with yours about costs/quality.
     
  148. Whether it is cheaper to replace a defective unit after a complaint than to strive for 100% quality at the end of the assembly line is debatable. However, the better quality programs are designed so that a problem is not caught at the end of the assembly line but rather at a much earlier stage in the process, and it is widely recognized that this does save money.
    By the way, some of the very large companies (Motorola, GE, Johnson and Johnson, Allied Signal, etc.) who have adopted six sigma programs are reporting savings typically measured in hundreds of millions to billions of dollars.
     
  149. Do you think Motorola is going to have a 3.4 in a million defect rate on the Xoom because of this program? From the limited information we have, Canon's defect rate is similar to others in their industry of lens making. Are you suggesting the entire Japanese lens making industry is ignorant of or unwilling to implement a reasonable statistical QC program? That a company the size of Canon would not have it or would not want to apply it to the lens making division?
     
  150. Erik, you said "Are you suggesting the entire Japanese lens making industry is ignorant of or unwilling to implement a reasonable statistical QC program?"
    Answer: Not quite. I am suggesting that, based on a likely defect rate of a few percent at the present time, the Japanese lens making industry is not at a high level of overall quality. A good statistical QC is an important part of it, but alone it is not sufficient. I do not know what quality program Canon has adopted, but whatever it is it is not presently producing a high level of quality if we assume it is allowing a defect rate of several percent.
    Earlier I gave results an example of a six sigma calculation showing that if a product has a parts count of 300 then a six sigma quality program will produce a defect rate of 0.1% in the final product. By a similar calculation, if the product has a parts count of 1000 then a six sigma program will produce a defect rate of 0.34%. A product with a parts count of 10,000 will have a six sigma defect rate of 3.4%.
    Based on the best evidence so far it appears that the defect rates for lenses are probably in the range of a few percent. I don't think the parts count of a lens is measured in the thousands, and if this is true then the current quality performance of lens makers is well under six sigma.
    I don't know what the parts count of a typical lens is. If anyone would like to present a number (hypothetical or real) it is not hard to calculate the defect rate for a six sigma process.
    In these calculations I am applying six sigma to the individual parts and not to the end product. If one were to apply it to the end product (which might not be realistic for lens production) the defect rate would be 3.4 in a million units produced.
    One other bit of info. A five sigma process has a defect rate of 233 DPMO, and a four sigma process has a defect rate of 6210 DPMO.
     
  151. I do not know what quality program Canon has adopted, but whatever it is it is not presently producing a high level of quality if we assume it is allowing a defect rate of several percent.​
    Now we are getting back to high/low compared to what? You never did come up a comparably complex and comparably priced consumer product that is significantly better. BTW, it's not just physical parts; assembly steps also add to complexity and defect opportunities. Could quality be higher? Undoubtedly. Would costs be higher? You originally claimed no, but now say it's debatable. Canon cameras/lenses may not be very innovative, but iterative incremental improvement is something they are known for. If there was a simple answer to the cost/quality equation, why would they ignore it?
     
  152. So Erik, do you think a ~2% defect rate is a good quality record for something you spent ~$1000 to buy?
     
  153. Spearhead

    Spearhead Moderator

    Speaking as a photographer, rather than my father's proofreader, why would I look at anyone's defect rate rather than whether or not I'm doing a good job as a photographer. Same thing applies to others. In the end, it's about the photographs. Looking at yours...whoops...hard to say. Manufacturer defect rates, whether real, or completely unvalidated like the ones here, aren't relevant to me as photographer or as a judge of other people's capabilities. I don't even know why anyone would anser that question with a yes or no.
     
  154. Erik,
    You asked for statistics on a consumer product of comparable price and complexity with a lower defect rate. Those statistics are hard to come by. However, I have already pointed to air travel, which is probably at least as complex as building a piece of photographic equipment.
    If you don't like the air travel example then how about a consumer product that probably is an order of magnitude more complex than a lens and has a 2% defect rate? Consider the Hyandai Sonata automobile.
    http://www.fordforums.com/f655/mustang-has-lowest-defect-rate-per-100-cars-all-american-cars-100368/
     
  155. Alan, when you did not answer my previous response I assume you had none. I'll repeat:
    Let's ignore it's a service and not product. What aspect of "quality" are you using: Arrive alive? Arrive on time? Arrive on time and satisfied with the service? You will get vastly different quality indices from these 3 different measures. What are the equivalent lens measures: Lens arrives in one piece? Lens performs within factory specifications? Lens performs within user expectations? You will get vastly different quality indices from these 3 different measures.
     
  156. So Erik, do you think a ~2% defect rate is a good quality record for something you spent ~$1000 to buy?​
    A 2% defect rate means nothing to me unless I get one of the 2%. So far, in the dozen EOS lenses + 3 EOS digital + 4 EOS film bodies I own, I never have (and yes, I've done the picky focus tests). As long as the issue can be solved in reasonable time under warranty or exchange, then it's part of the risk I take to get the item for $1000 from a convenient source.
    But then again, I really have no choice. Switching to another brand won't help. I spend as much or more on things that have higher defect rates, e.g. almost all consumer electronics, automobiles, etc.
    P.S. I take that back: I did get one of the early 70-300mm IS lenses from the serial number range that was recalled to fix a problem where the lens focus barrel may become loose over time. My lens operated just fine but I sent it back to Canon for warranty repair as a precaution.
     
  157. James Meketa , Apr 12, 2011; 12:41 a.m.
    Every time I return a defective lens to B&H or Adorama, I know that it will be repackaged and resold as new to someone who doesn't know any better.​
    And then:
    James Meketa , Apr 16, 2011; 08:41 p.m.
    My earlier post (April 12) about returns being recycled was based on what I was told explicitly by the owner of a large, east coast camera store when I asked him what would happen to a brand new, but badly decentered, prime lens I was returning. I assumed from his reply that this practice was standard. I am glad to learn that it may not be.​
    So, on Apr 12th you posted as a fact which you claimed to "know," something it turns out you presumed to be true based on unsubstantiated information you say you received from the owner of another (unnamed) camera store. First, you presume B&H and Adorama do what this other store apparently says they do. You can hardly be certain of this. Second, whoever this unnamed owner (of an unnamed store) is, he or she certainly considers B&H and Adorama his (or her) competitor. That hardly makes him or her someone who'd know what we do and certainly calls into question his or her unbridled objectivity.
    Henry Posner
    B&H Photo-Video
     
  158. ShunCheung

    ShunCheung Administrator

  159. Interesting thread on DPReview about two "defective" lenses in a row from B&H: http://forums.dpreview.com/forums/readflat.asp?forum=1030&thread=38261050
    And, as it turns out, neither lens was in any way defective. The OP of that thread simply didn't understand how his lens worked. So, for the first lens we shipped, which was not the least bit defective and was apparently in as-new condition when we got it back, what's an ethical retailer to do with that item?
    Henry Posner
    B&H Photo-Video
     
  160. Henry,
    I know you won't do it but as far as I am concerned that buyer has cost you money, I would bill him for both lenses, or as an absolute minimum the costs you will incur, as neither was defective, he should pay for his mistake. American retailing is a strange beast though, the cost of manipulative and inexperienced buyers has to be absorbed by everybody, yet very generous return policies are expected almost as a right.
    This illustrates my point perfectly, most people who return lenses as defective are actually wrong.
     
  161. "...most people who return lenses as defective are actually wrong."
    Another unsubstantiated opinion stated as if it were a known fact. Maybe the opinion is true, and maybe it isn't, but either way it is no more than an unsubstantiated opinion.
     
  162. Alan, taken out of context maybe, but I was referring to specific instance where even the returnee admits he was wrong!
    You can argue the figures and semantics endlessly, oh you are, but that doesn't change the fact that none of us know a true defective lens figure. One of my points has been that I believe many of the returned lenses that are declared defective and are reported as such, are not, that was shown to be true in this instance.
    You argue, constantly, that a possible 2% defect rate is far too high for you. For me it is fine, my life does not depend on the ability of a lens to take a sharp image out of the box, as opposed to a cars brakes, fuel system, steering etc or a planes landing. But cars are very far from reliable and many millions are recalled over direct safety issues, across many manufacturers. Where is your six sigma there?
    If I didn't have the ability to return a defective lens (as opposed to abusing a retailers return policy because I am an idiot) then maybe I would demand a better rate, but the manufacturers do stand behind their products and will exchange them if the originals are bad. No harm, no foul. If my plane crashes I won't be so alive, but when did I last have a flight I was content with at the end?
     
  163. I quote:
    **** "Henry Posner [​IMG], Apr 22, 2011; 09:50 a.m.
    Interesting thread on DPReview about two "defective" lenses in a row from B&H: http://forums.dpreview.com/forums/readflat.asp?forum=1030&thread=38261050
    And, as it turns out, neither lens was in any way defective. The OP of that thread simply didn't understand how his lens worked. So, for the first lens we shipped, which was not the least bit defective and was apparently in as-new condition when we got it back, what's an ethical retailer to do with that item?
    Henry Posner
    B&H Photo-Video**** "

    I presume your question may well be rhetorical. However, as it has been asked I think that there are several things that an ethical retailer can do:
    1. Sell the lens as used, class II or demo (whichever fits the situation as to the usual and customary business practices in your particular type of business in addition to acting correctly within the local and state laws) at a reduced price. Yes, you'd take some loss on top of the shipping/packaging/return item costs but sometimes things like that are just the price of doing business.
    2. Return to vendor (if your vendor contract allows).
    3. See #1 above but have your buyer negotiate with the vendor to eat some or all of the sales price difference between a new lens and the price you'd get selling under the #1 above conditions.
    4, 5, 6, etc... are actions that I'm sure you could add to the list having been in that business for some time.
    Although the errant customer should offer to offset your loss it is unlikely they will. I think it would be a business mistake for B&H to try to collect monies from him/her. The cost of collection would be more than your net loss on a lens. He/she probably feels pretty sheepish. If I were in your shoes, I'd have a B&H tenured sales type call to ask the customer if everything is OK now. Have the salesperson reassure him/her that these things happen and there are no hard feelings at B&H. I'd think you wouldn't want the customer to be too embarrassed to ever buy from you again. I'd also think you'd want the customer to tell their photo friends (many of whom are leery of dealing with eastern seaboard photo retailers) how nice, ethical and understanding B&H is.

     
  164. Alan, I don't know what quality statistics you are reading (or misreading), but quality in automobiles is typically expressed to as defects per 100 vehicles. Here is a blog from Hyundai in 2009 where they brag that their defect rate is 95 per hundred vehicles well below the industry average of 108.
    http://www.hyundai-blog.com/index.p...yundai-4th-in-its-2009-initial-quality-study/
    I strongly suspect that whatever source you are using has a very different measure of quality than the one used by most of the industry.
     
  165. Erik,
    In your last post you asked what quality statistics I was reading with respect to the Hyundai. No need to wonder about the source... all you had to do was follow the link provided in my earlier post and read. However, if you don't want to follow the link, here is a quote from it. Go the the fifth paragraph which starts "The 2004 Hyundai Sonata..." for the part about the defect rate in that model of automobile.
    "U.S. Automakers Improve Reliability
    3/7/05

    DETROIT (AP) -- U.S. automakers improved the reliability of their
    products last year, but Hyundai Motor Co. and other Asian companies
    still make the most trouble-free vehicles, according to a survey
    released Monday by Consumer Reports magazine.

    Customers reported an average of 17 problems per 100 vehicles for 2004
    models from DaimlerChrysler AG's Chrysler Group, Ford Motor Co. and
    General Motors Corp., the magazine said. That was down from 18 problems
    per 100 in 2003.

    Japanese and Korean automakers had a rate of 12 problems per 100
    vehicles -- unchanged in the magazine's last three surveys. European
    automakers, some of whom have battled quality issues in recent years,
    had 21 problems per 100 vehicles. That's up from 20 a year ago.

    The survey is part of Consumer Reports' annual auto issue, scheduled to
    hit newsstands Tuesday.

    The 2004 Hyundai Sonata was the most reliable vehicle in 2004, with two
    problems per 100 vehicles. Consumer Reports said the Sonata is "further
    establishing Hyundai's remarkable turnaround from one of the least
    reliable brands to one of the best."

    As an overall brand, Hyundai recorded a reliability rating of 11
    problems per 100 vehicles, tying it with Toyota Motor Corp.'s Lexus and
    Nissan Motor Co.'s Infiniti nameplates. Subaru was the most reliable
    brand in 2004, with an average of eight problems per 100 vehicles.

    Reliability can vary widely within a company. The 2004 Ford Mustang was
    the most reliable car made by a U.S. manufacturer, with five problems
    per 100 vehicles, the magazine said. But Ford's Lincoln Navigator sport
    utility vehicle tied with the Nissan Quest minivan as the least
    reliable, with 49 problems per 100 vehicles.

    Consumer Reports measures reliability by surveying its subscribers. The
    magazine collected data on a record 810,000 privately owned or leased
    vehicles, 20 percent more than the 675,000 vehicles included in last
    year's survey.

    The magazine asked subscribers to report serious problems such as faulty
    air conditioning, wind noise, electrical difficulties and transmission
    trouble.

    Also Monday, Consumer Reports announced it was no longer recommending
    the Ford Focus as a top pick among small cars after the Focus got a poor
    rating in side-impact crash tests performed by the insurance industry.
    The Insurance Institute for Highway Safety released new crash test
    results Sunday.

    Ford responded that the Focus got better side-impact ratings from the
    federal government, which also performs crash tests. The company said
    the Focus also did well in the reliability survey.

    "We recognize how important it is to make sustainable progress in
    quality, and we won't be satisfied until we are the best," Ford said in
    a statement.

    Consumer Reports also said it would no longer recommend six other
    vehicles because of the insurance institute's side-impact crash tests.
    Those vehicles are the Honda Element, Mitsubishi Outlander and Suzuki
    Grand Vitara SUVs, the Nissan Altima sedan and two small cars, the
    Hyundai Elantra and Mazda 3.

    Consumer Reports buys all the vehicles it tests and doesn't accept
    advertising."
     
  166. Two thoughts:
    1. I don't think I've visited this thread in over a week now, but I just saw that it contains over 160 posts, so I decided to see what the heck might warrant that. Went straight the last post to find that...
    ... we are now ranting about the reliability of cars!? What the...!
    2. Then I looked a bit higher on the page and saw this:
    Henry Posner wrote:
    And, as it turns out, neither lens was in any way defective. The OP of that thread simply didn't understand how his lens worked. So, for the first lens we shipped, which was not the least bit defective and was apparently in as-new condition when we got it back, what's an ethical retailer to do with that item?
    That is a very important observation, and it is in line with quite a bit of what I read in photography forums in general when people start commenting on the quality of equipment and making gross generalizations based on... based on what, actually?
    I'm convinced that quite a bit - the majority most likely - of the "facts" about gear offered in forums are really far from facts. The non-facts fall into several interesting categories:
    • Naively and uncritically repeating stuff read or heard somewhere else that passes for "wisdom." There are so many myths and so much silliness passed off as factual that you should take all forum posts (including mine! :) with a grain of salt. (Disclaimer: Not all posts are nonsense, but be careful out there!)
    • Buying into hyperventilating complaints about this or that piece of gear. In order to not divert the thread too far, I won't identify the lens, but I often hear that a particular lens is "soft" in a certain type of use. I own the lens and use it a lot and I produce really sharp prints from it in this "type of use" at quite large sizes. In other words the frequently repeated "fact" about this lens is actually utter nonsense.
    • Not knowing how to use the gear correctly or simply not wanting to bother using it correctly. If I had a nickel for every "my lens isn't sharp" post from someone who made a hand held shot in low light without paying attention to where the AF points were and used a shutter speed that was too low, I could buy some really cool gear! (Or more ink for my printer! ;-)
    • Related to the above, assuming that "I spent a lot of money on this, so it should make perfect photos" regardless of how I use it. This is sort of like being an inexperienced driver, getting out of a ten-year-old beater Corolla, stepping into the most expensive BMW... and being really upset because I still run into the garbage cans when I back up without looking. (If I also had a nickel for each of these, I'd be verging on wealthy.)
    • Obsessing about imaginary "perfection" in photographic equipment. We all know the type - this person isn't really interested in photography (much less in photographs) as much as in the quest to possess objects of technical perfection. Such people almost invariably expend inordinate effort searching for the tiniest and least significant marginal differences - differences that are completely inconsequential in these things called photographs that we hope they someday discover - and instead send back equipment because when they stare at their screens at 100% magnification long enough and flip obsessively and repeatedly between two samples of something they think they can, if the look really closely, maybe detect that the two are not perfectly identical and that, therefore, one must be good and the other "not good."
    In some ways it must be crazy selling camera gear to us.
    Dan
     
  167. Dan,
    The reason cars were brought into the discussion is because Erik Magnuson requested information on the defect rate of other products besides lenses.
     
  168. Alan, a link with a selected quote would have been good enough rather than posting the whole article. However, this goes back to the point I tried to make when you brought up air travel: what exactly is a defect? Different definitions will give different defect rates. For example, is a single car with multiple problems one defect or many? The strict definition used by the industry shows a much higher defect rate than self reported incidents in a magazine survey.
    Also you are forgetting one big difference: dealer preparation. A car dealer is supposed to carefully inspect each car before it is sold and fix any factory defects. They usually add a few hundred to a few thousand dollars to the sticker for this inspection. I'm sure you can find a camera dealer who would be willing to provide the same service for a similar price (i.e. selling you the lens at list price instead of MAP.)
     
  169. "...most people who return lenses as defective are actually wrong."​
    Maybe "most" was hyperbolic but I linked to a customer who got two lenses from us and only after the second exhibited the same "defective" behavior did he bother to ask whereupon he learned neither was defective at all. It happens too frequently.
    An anecdote from the film era. A customer emailed me while returning the third "defective" Nikon FM3a body. He was upgrading from a simpler consumer Nikon slr. On that camera, opening the film back meant simply lifting the rewind knob. On the new FM3a, like the FM and FE series (and several other pro-oriented Nikon cameras) you had to actuate a small lever before lifting the rewind knob. He didn't know this and didn't bother to read the manual, so we shipped and got back three perfectly perfect cameras.
    Maybe saying "...most people who return lenses as defective are actually wrong," is an exaggeration but I'd stand by saying "...too many people who return lenses as defective are actually wrong."
    YMMV
    Henry Posner
    B&H Photo-Video
     
  170. Henry,
    Only you know the figures, but if your return rate for lenses across Canon, NIkon, Sigma and Zeiss is more than 2% I'd be thinking all of the extras are not defective. If you are getting 4% then 50% of the returns are probably unjustified, if the figure is anything like the 8% hypothesised earlier in the thread then 75% are probably wrong.
    I'd love the savings of not having to absorb the overly generous return policy to be carried over to purchasers who know about rewind knobs and macro dof notifications :)
    Take care, Scott.
    P.S. Is that really cool stainless steel diner still across the street from you?
     
  171. P.S. Is that really cool stainless steel diner still across the street from you?​
    The very nice Skylight Diner's on the south-west corner of W34th St & 9th Ave. The roach-trap a block south on W33rd St & 9th Av has been deservedly demolished.
    Henry Posner
    B&H Photo-Video
     
  172. It's been a few years since anyone has posted on this thread. While the thread was more active someone asked for examples where six sigma quality had been achieved. Well the organization where I work has recently achieved six sigma performance in an important measure of quality for one aspect of its operation, namely mislabeled or lost units. This is an operation that handles tens of thousands of units per day received from hundreds of different sources each day.
    Also, it can safely be said that the overall defect rate per completed unit is WELL under one per thousand units. I can't give you more specific information on that figure because it could be considered business sensitive information, but it is very likely that competitive organizations have roughly comparable defect rates to ours. Otherwise their businesses are likely at risk.
     

Share This Page