Are the lemons just miscollimated cherries? (1 Viewer)

[WJC]

While I have already offered my two-cents worth on this thread, I would like to offer a nickel more.

I have not read the Cornell review in question. However, I can say without a doubt that it is perfect. How do I know that? Because these people, no matter who they are, have a right to offer their opinions—based on whatever—on anything they want.

The problem, as I see it, is in what we opto-geeks expect of our reviews. At the risk of being redundant, being a master birder does not qualify one to know any more about the more technical side of binoculars any more than being a disk jockey qualifies one to know the exact specs on the microphones they use. The microphones and binoculars are just tools to those who use them.

The level of opto-geekolgy needed to please folks like us cannot come from Cornell, but rather the University of Arizona’s Optical Science Center, the University of Rochester’s Institute of Optics, Lawrence Livermore Laboratory, or other such institutions around the world. This is so, because some of the data sought and fought over has nothing to do with birding but, rather, hard mathematics and optical engineering—not to mention a knowledge of the binocular industry.

From my standpoint, the level of info sought by many on the list is a MASSIVE overkill to what is actually needed to make a good buying decision. However, most of what I have learned has been learned the hard way. So . . .

Just a thought.

Bill

 

[kabsetz]

[Of the three basic issues mentioned, the most difficult to deal with is clearly the products' quality distributions. Let me first clarify, though, that my suggestion to use a random sample size of N=3 was not to get the best estimate of this distribution. It was presented as an alternative to pre-selecting the best single examples, while minimizing the possibility that a given manufacturer might be hung out to dry with a single out-of-tolerance specimen if N=1. In principle, the enigma with stopping the search for specimens when the evaluator's selection criterion is met is that Company A might have great quality contol and qualify on the first try, whereas Company B, with poor QC, might take three or four tries to get there. Indeed, the run length required itself reflects the quality distribution quite independent of how the final examples ultimately compare. If all members of a fixed sample size were measured and tested, however, over time it would be possible to correlate aggregated star-testing results with the other evaluations using statistical methods. (Until done who knows the results?)

Yes, I think your suggestion makes sense. Deciding that three is the number of units for each model to be screened and then sticking to that number is a practice which may strike a practical compromise between the many issues involved. To stop the search at three obviously also saves time and energy over a non-limited search. However, it might not always be possible to get to screen more than one sample, which is why I have taken the easier route of accepting the first sample if it is not obviously faulty. When the first sample does show obvious flaws and publication deadline and other factors make it impossible to evaluate more samples, I have indicated my suspicion of a possibly subpar sample in the text. For a reader, though, it would be more valuable if there would always be a fixed number and it would be indicated, even if on very approximate terms, how samples A and B related to the tested sample C.


[Getting back to the difficulty of measuring the product's real quality distribution, my feeling is that there are simply too many influences potentially at work, many of which you've already mentioned, and very likely major differences is each company's sampling inspection plan(s) and resulting operating characteristic curve. Basically, these determine the tradeoffs of so-called producer and consumer risk, but there is no law of man or nature that requires each company to have the same business strategy. A company that samples from very small production batches and uses narrow criteria will obviously market a more consistent product but at a higher cost (all else equal) — which must be reflected in the profit or price.]

And since we don't know whether high price equals high profit or improved quality control, and since (I suspect, anyway) that good quality control is not only a matter of investment and procedure, but also a product of the qualities of the people implementing it, we really cannot tell how things stand without an extensive survey.

[So, putting it all together, my advise would be to forego the measurement of quality distributions (which involves estimating the distribution's shape/form) and simply stick with a fixed sample size that does not exceed testing resources, but does give the consumer some modicum of insight into the product's variation. Right now, simply including the run length and stats for rejected samples would be of value if not done already.]

Yes, this sounds eminently sensible to me.

And finally, if we were to try to get a grip on real quality distributions, I feel the only way to approach it would be to have a continuous consumer survey (much as is being done with motor vehicles in Germany and probably elsewhere). In such a survey, users using a simple and standardized star-testing procedure would provide key data about the units they own. However, my head immediately begins to hurt when I even scrape the surface of the potential distortions and pitfalls such an approach would contain.

[I hope this hasn't worked out to be too arrogant a post, and as usual it only reflects my peculiar opinions.]

Don't worry, I did not find anything arrogant in your post, although perhaps I did not do a close enough reading between the lines.

And finally, I would like to add that although it would be very interesting and even valuable to know precisely what the quality distributions for various manufacturers and products are, I try to keep in mind that the goal (for me, anyway) is that a birder in the field have a good piece of optics in their hand. Thus, it helps only a little to know that with brand (A) I may have a 50% chance of getting a reasonably aberration-free specimen while with brand (B) that probability is 65%. I certainly would not stop at knowing that by choosing B over A my chances are somewhat better. The important thing for a birder is to know that variation exists and that the responsibility for making sure that one gets a good sample ultimately rests on one's own shoulders.

Kimmo

 

[elkcub]

Don't worry, I did not find anything arrogant in your post, although perhaps I did not do a close enough reading between the lines.

Kimmo,

Hey, I have a hard enough time writing what's on the lines much less putting anything between them.

It would appear that we basically see eye to eye, and I'd like to thank you for a very worthwhile discussion. In future I hope we can look forward to discussing the specifics of a double blind study to establish the extent to which star testing results might be related to normal birding perceptions. That should probably go on a new thread.

Regards,
-elk
PS. It's a rare person who would even consider making changes to a well-established way of doing things. My compliments.

 

[iporali]

Very nice discussion - thank you all! There have been some great "concluding" posts to this thread, but I still would like to ask Kimmo, Bill or any others something that I haven't seen answered yet - sorry about that

What can you do if you have purchased an average (or slightly less, not necessary a lemon) sample, are unhappy with it and are willing to pay anything to make it a cherry? Is it more often a matter of component or assembly quality?

Of course one solution is to sell the unsatisfactory sample on e-bay and buy another until you find a unit you are happy with. But if good optical repairers can make broken binoculars good again, why wouldn't they be able to assemble a pair of binoculars better than someone who does not have time to put all his/her effort to it.

Best regards,

Ilkka

 

[Pileatus]

Very nice discussion - thank you all! There have been some great "concluding" posts to this thread, but I still would like to ask Kimmo, Bill or any others something that I haven't seen answered yet - sorry about that

What can you do if you have purchased an average (or slightly less, not necessary a lemon) sample, are unhappy with it and are willing to pay anything to make it a cherry? Is it more often a matter of component or assembly quality?

Of course one solution is to sell the unsatisfactory sample on e-bay and buy another until you find a unit you are happy with. But if good optical repairers can make broken binoculars good again, why wouldn't they be able to assemble a pair of binoculars better than someone who does not have time to put all his/her effort to it.

Best regards,

Ilkka

Ilkka,

I'll bet that most warranties are null and void the moment someone other than the manufacturer gets inside them. Bill knows the repair business so maybe he can clarify.

Do you think you have a sub-standard model?

John

 

[iporali]

Do you think you have a sub-standard model?

Oh, no - I am perfectly happy with our Nikons (10x32HG & 8x32SE)... honest!

Ilkka

 

[WJC]

Ilkka,

I'll bet that most warranties are null and void the moment someone other than the manufacturer gets inside them. Bill knows the repair business so maybe he can clarify.

Do you think you have a sub-standard model?

John

Buy a good bino and be happy.

My garden has lots of rocks in it. Conglomerate, gneiss, scoria, some good granite, and a few pieces of quartz. I would much rather they be diamonds, opals, and rubies. Alas, no.

To make a cheap bino into an expensive binocular would cost 20 times as much as buying a high-quality, expensive bino to start with . . . and waste HUGE amounts of time. The reasons for this are way too numerous to address in detail.

If there are any specific questions about the process, I would be pleased to give a stab at answering.

Cheers,

Bill

 

[Leif]

I would guess that the only improvement an optical technician could make would be to improve the collimation, assuming that it was not up to scratch. However, I'm not sure if this can be done for roof prism bins without access to the manufacturers facilities. Unless I am mistaken when Swarovski repair an instrument they cut off the rubber armour and discard it to gain access to the frame, then put on new armour when all is kosher. Other factors for a lemon could be poor coatings on one or more surfaces, and optical surfaces not to quality. I doubt a repairer could do much for that.

Leif

 

[elkcub]

Very nice discussion - thank you all! There have been some great "concluding" posts to this thread, but I still would like to ask Kimmo, Bill or any others something that I haven't seen answered yet - sorry about that

What can you do if you have purchased an average (or slightly less, not necessary a lemon) sample, are unhappy with it and are willing to pay anything to make it a cherry? Is it more often a matter of component or assembly quality?

Of course one solution is to sell the unsatisfactory sample on e-bay and buy another until you find a unit you are happy with. But if good optical repairers can make broken binoculars good again, why wouldn't they be able to assemble a pair of binoculars better than someone who does not have time to put all his/her effort to it.

Best regards,

Ilkka

Ilkka,

Personally, I'd run any high-end company ragged before selling and repurchasing the same model again. After all, it's the company's QC policy that gave you the aggravation in the first place. If you know they can do better, it just ain't the time for restraint. What could you lose?

Best regards,
-elk