A student considering switching Nikon for Canon.

[inTempus]

I thought I would investigate optical decentering as mentioned several times by Derrel now regarding Canon. His comments seem to imply that Nikon has perfect lenses, or nearly perfect, while Canon lenses are plagued with issues. Derrel has also implied that Nikon tests all of their lenses whereas Canon doesn't, and thus Nikon warrants the premium they charge for their lenses. I would like to see a source for this claim as I can't seem to find anything definitive on my own.

First, optical decentering isn't an issue encountered with Canon only. It happens with all manufacturers, Nikon included.

Here's an article that touches upon the issue and cites a specific Nikon lens this reviewer had the issue with.

Lens Evaluations

An issue frequently left out from reviews is the variation found in optical products. All manufacturers gloss over this and say that their products are "quality tested" before shipping. Take it with a pinch of salt. This statement is virtually always true for long lenses, often true for medium long lenses, and more commonly false than you would like to care about for short lenses and zooms.

...

A typical example is the exciting AFS 17-35 mm f/2.8 Nikkor, which incidentally is a lens that I literally had available by dozens. I found in early production runs of this very expensive lens samples that showed mild degrees of optical decentering, or focus shifts within the zooming range.

It doesn't seem to be something you can find a lot of info on with a Google search.

But I did find an interesting survey conducted by Lens Play.

Survey Results

Note: Values are greyed out if there are less than 300 lenses sampled


Canon lenses - 11393 with 853 defects
The probability of getting a good lens is 93 %
The probability of getting 5 good lenses in a row is 68 %


Sigma lenses - 2405 with 545 defects
The probability of getting a good lens is 77 %
The probability of getting 5 good lenses in a row is 28 %


Tamron lenses - 1203 with 183 defects
The probability of getting a good lens is 85 %
The probability of getting 5 good lenses in a row is 44 %


Tokina lenses - 404 with 67 defects
The probability of getting a good lens is 83 %
The probability of getting 5 good lenses in a row is 40 %


Nikon lenses - 2038 with 163 defects
The probability of getting a good lens is 92 %
The probability of getting 5 good lenses in a row is 66 %

It would seem, based on their survey findings, that statistically you're more likely to get a good Canon Lens than a good Nikon lens.

Would I put much stock in this? Not really. I would say Canon and Nikon are pretty evenly matched and a few percentage points here and there aren't to be taken too literally.

What it does show is that the notion you're more likely to get a crappy Canon lens vs. a Nikon lens is more than likely false.

I am open to other surveys or studies that show otherwise though and would welcome them.

 

[KmH]

Statistically speaking, the list you posted is useless because the sampling rates are seriously skewed. Note there were well 11,400 Canon lens responses compared to 2050 responses concerning Nikon lenses.

One might want to explore, why the nearly 6 to 1 response rate? Something seems to be out of whack, relative to the market share of the 2 companies.

 

[inTempus]

Statistically speaking, the list you posted is useless because the sampling rates are seriously skewed. Note there were well 11,400 Canon lens responses compared to 2050 responses concerning Nikon lenses.

One might want to explore, why the nearly 6 to 1 response rate? Something seems to be out of whack, relative to the market share of the 2 companies.

While I appreciate your comments, I don't believe I represented the survey as a scientific study. It's also not "useless" as you claim. It's far more useful than making a proclamation about the quality of a product without citing any source what-so-ever. A non-scientific study is far more valuable than no study. However, if you have something more definitive, please share it.

I would say the sample size for Canon is quite notable. 11,400 responses giving them a 93% rating isn't anything to sneeze at. Heck, even 2050 isn't a small number given organizations such as Gallup can predict elections by polling 1,000 people or less.

 

[chyidean]

I would say the sample size for Canon is quite notable. 11,400 responses giving them a 93% rating isn't anything to sneeze at. Heck, even 2050 isn't a small number given organizations such as Gallup can predict elections by polling 1,000 people or less.

I can see where you're coming from, and I see what you mean, but I just want to point out that the amount of people responding to an optional survey doesn't always indicate the quality of the responses.

A large sample size may be beneficial to render outliers insignificant but it cannot be used instead of proper sampling methods (ie random sample, proper representation of all members of the population, etc.) I think the percentages for ALL of them might be slightly on the low side, because people who have defective lenses are more likely to complain, resulting in a perceived higher reporting of defective lenses. People who have non-defective lenses expect the lenses to be not defective so they're less likely to have a strong opinion and respond to an optional study.

 

[Derrel]

Here's an example of a Canon "professional, L-model lens" that displays poor quality control from Canon. This is from photozone.de, one if the web's larger, independent lens testing web sites.

Canon EF 24-70mm f/2.8 USM L - Test Report / Review
MTF (resolution)Well, I guess everybody has a nemesis and mine is the Canon EF 24-70mm f/2.8 USM L. It took me 4 (f-o-u-r) samples of the lens to get a good one - please note: "good", not a "great" sample. The first three variants showed rather hefty centering defects which spoiled the results quite a bit.This final sample exhibited a very good to excellent center resolution at wide-open aperture throughout the tested focal length range. The borders follow on a good to very good level. Stopping down lifts the center further into excellent territories and the borders improve gradually till about f/5.6 to f/8. The sweet spot of the lens is in the middle of the zoom range.

And Canon's 70-200mm f/2.8 L-IS USM lens?
The third time was the charm--the lens was tested on its third try... Here is the URL and the quote. Poor Klaus, trying to get decent samples to test...
Canon EF 70-200mm f/2.8 USM L IS - Test Report / Review
"MTF (resolution)
Two sample of this lens has been tested. The first sample performed quite poor and the 2nd sample did outperform its quality figures but not without problems. This 2nd sample has been tested once more after a calibration by the Canon service.
In the lab the lens showed a very good (f/2.8) to excellent performance at 70mm and 135mm However, at 200mm the resolution figures are not all that impressive for such a high-priced product. The figures are generally still very good here but frankly you can have a better performance out there. Regarding the fact that the lens has been calibrated it is viable to state that the lens is not any better than that. Regarding the quality figures at 200mm it does not seem to be advisable to use this lens with converters without hefty performance penalties."

Again, pretty unimpressive performance in terms of Quality Control for a professionally-priced, L-series zoom lens that would be counted on to be a cornerstone of any serious shooter's lens kit.

How about another Canon, L-glass, high-priced "professional zoom"? Canon EF 16-35mm f/2.8 USM L - Lab Test Report / Review
"The lens exhibited a very good but slightly varying performance in the lab. The center performance is excellent throughout the zoom range with a superb peak at 24mm. Unfortunately the border quality isn't quite up to that level specifically at large aperture settings at the wide and long end of the zoom range. The extreme borders at 16mm (not shown in the graph below) suffered from field curvature effects. As usual the quality can be increased by stopping down resulting in very good (16mm and 35mm) to even excellent (24mm) levels.
Two samples of the lens were tested during the recent months. The first one tested in late '05 showed slightly better figures at 35mm and slightly worse at 24mm. However, in the field the results were somewhat sub-standard under certain conditions so the lab results were not taken into account here despite a statement by the Canon Service that the lens was within specs."

 

[inTempus]

I can see where you're coming from, and I see what you mean, but I just want to point out that the amount of people responding to an optional survey doesn't always indicate the quality of the responses.

Yes, once again - this isn't a scientific study. It's the only thing I can find on the internet that's unbiased and deals with the quality of lenses purchased by owners of all major manufacturers. Again, if you have something better, please share it.

The best possible solution would be for the manufacturers to give us their repair data, but good luck getting any of them to release that info.

A large sample size may be beneficial to render outliers insignificant but it cannot be used instead of proper sampling methods (ie random sample, proper representation of all members of the population, etc.) I think the percentages for ALL of them might be slightly on the low side, because people who have defective lenses are more likely to complain, resulting in a perceived higher reporting of defective lenses. People who have non-defective lenses expect the lenses to be not defective so they're less likely to have a strong opinion and respond to an optional study.

Slightly on the low side? Perhaps. Maybe their happy customer rates are higher than 93%. Maybe is 95% or 98%. The point is, it's still pretty close if we assume your assumption is accurate. Most properly conducted polls have a margin of error in the 3-5 point range. That means this survey is probably pretty solid.

For the purposes of this discussion, let's assume the number for both Nikon and Canon are really 98%. That means the notion Canon produces inferior lenses while Nikon produces perfect lenses is still inaccurate. After all, that's what this is all about, right?

 

[lbridges]

...
A large sample size may be beneficial to render outliers insignificant but it cannot be used instead of proper sampling methods (ie random sample, proper representation of all members of the population, etc.) I think the percentages for ALL of them might be slightly on the low side, because people who have defective lenses are more likely to complain, resulting in a perceived higher reporting of defective lenses. People who have non-defective lenses expect the lenses to be not defective so they're less likely to have a strong opinion and respond to an optional study.

I agree whole-heartedly about population representation. However, and I do note the numbers were presented as not being scientifically derived, the scaling factor for dissatisfied consumers would apply equally assuming the data collector did not purposely induce skewed data.

Here is a web reference based on numbers I would expect to be without bias (but have no evidence that they are/are not): LensRentals.com - Lens Repair Data 2.0

 

[Derrel]

A couple of comments to the OP. Why do you suppose somebody who actually has lens testing equipment and know-how went through so many samples of two popular professional lenses, in order to get a "professional" 24-70 wothy of representing its brand?

And why did it take a three tries to get a 70-200-L that was worthy of representing its brand; and why were the results of the professional 16-35 and the 70-200 "questioned" by a qualified lens tester. Why was Canon's assurance that the lenses, plural, were within specification questioned by an experienced lens tester? Why was there doubt that the lenses, after factory service, were really "good enough" to represent their model and their brand?

Another question for the OP: why do you suppose Nikkor lenses come with a 5-year warranty, while the white lens brand lenses come with a ONE-year warranty? There's a concept of standing behind one's product,and in this day of 90-day and 1-year warranty offers, Nikon's FIVE YEAR warranty is a statement in and of itself--the company warrants the fitness and reliability of its lenses for five YEARS.

As far as sample sizes and qualifications to judge: according to millions world-wide, McDonald's is *the* best restaurant in the world. Huge sample size, low price, tremendous satisfaction. Again, this is why I am citing photozone.de as a source--a *qualified and experienced* lens tester's opinion, to me, is worth a lot, while the opinions of a huge sample size of users unable to even discern if the lens is not all it could be, is of little value to me. Again, McDonald's....Ernest and Julio Gallo jug wine, etc.

5 year warranty vs a ONE year warranty. Higher cost? Not "always"--look at the Nikon 105 VR macro vs the Canon 105 IS macro; the Nikon is about $200 less costly, yet has a better warranty.

 

[inTempus]

A couple of comments to the OP. Why do you suppose somebody who actually has lens testing equipment and know-how went through so many samples of two popular professional lenses, in order to get a "professional" 24-70 wothy of representing its brand?

And why did it take a three tries to get a 70-200-L that was worthy of representing its brand; and why were the results of the professional 16-35 and the 70-200 "questioned" by a qualified lens tester. Why was Canon's assurance that the lenses, plural, were within specification questioned by an experienced lens tester? Why was there doubt that the lenses, after factory service, were really "good enough" to represent their model and their brand?

Another question for the OP: why do you suppose Nikkor lenses come with a 5-year warranty, while the white lens brand lenses come with a 3-year warranty? There's a concept of standing behind one's product,and in this day of 90-day and 1-year warranty offers, Nikon's FIVE YEAR warranty is a statement in and of itself--the company warrants the fitness and reliability of its lenses for five YEARS.

As far as sample sizes and qualifications to judge: according to millions world-wide, McDonald's is *the* best restaurant in the world. Huge sample size, low price, tremendous satisfaction. Again, this is why I am citing photozone.de as a source--a *qualified and experienced* lens tester's opinion, to me, is worth a lot, while the opinions of a huge sample size of users unable to even discern if the lens is not all it could be, is of little value to me. Again, McDonald's....Ernest and Julio Gallo jug wine, etc.

5 year warranty vs 3 year warranty. Higher cost ?Not "always"--look at the Nikon 105 VR macro vs the Canon 105 IS macro; the Nikon is about $200 less costly, yet has a better warranty.

All great commentary, but it doesn't answer the questions asked.

1) Where do you come up with your conclusion that Nikon has higher quality control than Canon with regards to lenses? Do you have any data to support his notion other than anecdotal evidence offered by two bloggers? I've shown actual data collected from thousands of owners which seems to counter your claims.

2) You've brought up optical decentering several times in recent threads, including this one, and act as though it's something unique to Canon... much like you acted as though micro-adjustment was developed just for Canon to deal with faulty lenses in this thread - and Nikon was immune. Obviously this isn't accurate either.

3) You originally claimed that Nikon cost more because they have something Canon lacks - "quality control".

Here's an example of why Nikon's higher-grade lenses cost more. Quality and quality control.

But again, this seems to be proven to be inaccurate given the information posted above. Now you're claiming it's because of a 3 year vs. a 5 year warranty?

I'm just looking for something more than unsubstantiated opinions. I'm willing to listen to any evidence you have to support your claims, but it seems you read something posted by a potentially biased source that has a very small sample set to collect data (one or two lenses) and you hold it up as proof positive Nikon rocks and Canon absolutely sucks.

 

[Derrel]

My mistake--I was crediting Canon with offering a 3-year warranty period--but I went to B&H Photo and Adorama, and I see now Canon's warranty is only ONE-year, so I amended my above post to reflect the 1-year warranty Canon offers, versus Nikon's 5-year warranty. And Tim, please stop making stuff up--like "Nikon rocks and Canon sucks"-that is YOU talking smack. You're a relatively new shooter, with a year's experience in serious photography; I myself recall my college days and even then in the early 1980's Nikon lenses were priced higher than Canon; I payed $269.95 for a 105mm f/2.5 Ai-S Nikkor in 1983, when minimum wage was $3.25 an hour, with a $2.40 hr take-home after tax rate: I payed for that lens with 112 hours of work. The lens works perfectly to this DAY. That same Nikon lens works on my current Nikon bodies AND on my current Canon bodies. Please note the lensrentals.com data as being representative of rigorous, demanding usage, not lenses sitting around in lens lockers. I differentiate between "cheaper" and "less-expensive". Why are there no Nikkor primes in the 16 Most-repaired list?

I also stopped by the lensrentals.com site and noted a few issues for the OP, who claims Nikon lenses COST more than Canon lenses, which would be the reason the OP was considering switching FROM Nikon to Canon--to get "cheaper" lenses. And please, keep in mind, I *own* the Canon 18-55, Canon 135 Soft Focus, Canon 50/1.4, Canon 85-1.8, Canon 135-f/2-L, Canon 24-105 f/4 L-IS USM, and the Canon 70-200 f/2.8 L-IS USM,and a 20D and a 5D, in addition to my full Nikon system which I have been amassing since 1982. I am not just a Canon-basher NOR a Canon fan-boy.

LensRentals.com - Lens Repair Data 2.0

Of the 16 Most Repaired Lenses, the worst five are Sigma zooms, with annual repair percentage figures from 84.6% down to 22.2%, followed by the Tamron 70-200 in sixth place at 22.2%. Seventh and 8th worst are a pair of Canon zooms, the fairly expensive EF-S 17-55 f/2.8 IS at 19.6% for IS and Err99 problems, so just under ONE-FIFTH of those requires repair each year. Keep in mind that a 17-55 f/2.8 IS-USM Canon is priced at $990, after a $70 rebate has been deducted, and sells with a ONE-YEAR warranty. According to the folks at lensrentals.com, 19.6% of their multiple 17-55's have been repaired on an annual basis. 17-55 would be a *critical* lens for wedding work on APS-C--but one-fifth per year FAIL??? WTF?

In eighth place is the Canon 10-22 EF-S which has a 15.8% annual failure rate for "barrel separation." Which is amusing, since a fellow TPF forum member here was bragging to me about how that lens, the 10-22 is "L-grade", but then about a week later, he told me his lens broke in two, so the design is obviously not too good. THis Canon F-S lens is $720 after rebate right now, and comes with Canon's now-standard ONE-YEAR warranty.

Nikon's first entry into the list comes with its 70-200 VR, the first generation model at ninth place and 13.1 percent annual repair rate.

In 12th position in the Most Repaired Lenses is Canon's 50mm f/1.2-L with an annual repair percentage of 13.3%-- the worst repair percentage per annual basis of any prime lens in their data. That 50mm-L glass lens sells for $1599 at B&H with ONE-YEAR warranty. Nikon has a new 50mm f/1.4 AF-S G lens selling for $449 at B&H Photo, and it comes with a FIVE-YEAR warranty.

Canon's 85mm f/1.2-L is in the 16 Most Repaired Lenses, with a one-year annual repair percentage 11.7%. The Canon 85mm 1.2-L sells for $1840 at Adorama with ONE-YEAR warranty.
Nikon's 85mm f/1.4 AF-D is priced at $1229 at Adorama and is sold with a FIVE-YEAR warranty.

Of the 16 Most Repaired lenses, there are three prime lenses. The most-likely prime lens to need repairs is Canon's 50mm f/1.2-L, followed by the Sigma 30mm f/1.4 EX, and then the Canon 85mm 1.2-L. No other prime lenses, from any maker, are among the 16 Most Repaired lenses. I find it surprising that two of Canon's newest, and highest-priced non-supertelephoto prime L-glass lenses make this list. Lensrentals.com's web page states they had 2,315 copies of 320 different lenses when this data was collected and presented.

Hmm...Sigma has 7 of the Most Repaired lenses; Canon makes 5 of the most-repaired lenses; Nikon makes 3 of the most-repaired lenses. Nikon offers a full FIVE-YEAR warranty on all the lenses listed, Canon offers a ONE-YEAR warranty on all the lenses listed.

 

[inTempus]

Lens Rentals data is a great resource. But a few things;

1) They only list the lenses with a 10% or higher repair rate. That means there are likely a whole bunch of repairs going unreported. I would like to see all of the data.

2) to make the list, they need to have 9 (and odd number) on hand. That means they probably only have 10 or so of any given lens. This isn't a very large sample set.

3) These lenses are treated like rental cars. I can assure you 99% of the owners of these lenses that aren't rental companies don't treat their property like, well, rental equipment. God only knows what these lenses are subjected to.

4) Even Lens Rentals cites the Lens Play survey. Interesting, because unlike Lens Rentals data, this is collected from thousands of owners vs. a rental company.

Now, with the rather small sample set and not having the full data set (lenses with repairs under 10%) saying Canon makes the top 5 and Nikon makes the top 3 is hardly saying much.

But let's get back to the whole quality control issue that keeps getting dodged. Where do you get your information that Canon suffers from optically decentered lenses and Nikon doesn't? Where do you get your information that Canon lacks quality control yet Nikon excels at it? Do you have anything to support these claims?

 

[inTempus]

I am not just a Canon-basher NOR a Canon fan-boy.

Have you read your own posts lately? You spend an inordinate amount of time writing novels about why Canon is bad and Nikon is great. It's not occasionally, it's daily. It's not a couple of sentences, it's paragraphs.

You obviously think you're unbiased, but honestly - take a look at the stuff you write. It's not the slightest bit objective, you're clearly on a mission and get quite worked up about it.

Just a friendly observation. I'm passionate about things too, but you seem somewhat obsessed with convincing people Canon sucks.

 

[lbridges]

...why do you suppose Nikkor lenses come with a 5-year warranty, while the white lens brand lenses come with a ONE-year warranty? ....

I presume you would therefore pick Hyundai/Kia over Honda/Toyota in terms of reliability, and indeed over even Acura/Lexus?

But since you asked in some serious sense -- I suppose it's offered as a marketing ploy to entice a consumer to purchase the product.

Do you honestly think if any company had a lock on the market they would give you anything more than a warranty that expired once you left the parking lot?

 

[PhotoXopher]

Do you honestly think that a company would warranty a product for 5 years if they thought they would lose money on the deal?

 

[Derrel]

I am not just a Canon-basher NOR a Canon fan-boy.

Have you read your own posts lately? You spend an inordinate amount of time writing novels about why Canon is bad and Nikon is great. It's not occasionally, it's daily. It's not a couple of sentences, it's paragraphs.

You obviously think you're unbiased, but honestly - take a look at the stuff you write. It's not the slightest bit objective, you're clearly on a mission and get quite worked up about it.

Just a friendly observation. I'm passionate about things too, but you seem somewhat obsessed with convincing people Canon sucks.

Sorry Tim, but you're the one on a mission...you keep resorting to fanboy type comments like "Nikon rocks, Canon sucks," and trying to attribute such comments to me. You referring to one of the web's largest lens testing sites a "blogger",etc,etc. is indicative of your ad-man style of put-down ad hominems in an effort to defend your beloved Canon brand. You are even the web master of Kwanonians, a Canon fan site. Your own signature, with your list of newly-bought L-glass lenses is quite a sign of gear-centric fanboyism. I love how you put the "L" in red type face! Sorry dude, I'm commtited to both brands, Canon and Nikon, for the forseeable future. I will use whatever is the better product to get what I need,and I shoot a mix of Canon and Nikon equipment. You? You've been at this whole game for a year,and are alresdy the web master at a Canon fan site??? Sorry buddy, but your sense of objectivity and your inability to listen to anything sombody with more experience says speaks volumes...

I OWN and USE a reasonably complete Canon system. YOU are the one who keeps using the word "sucks". Your juvenile approach is tiresome.

What's the URL for Kwanonians again??? I see you dropped the HUGE "Kwanonians" hot-link from your signature recently. Did TPF management ask you to drop that hot link, or did you just do it to appear impartial when you launched your "another unfavorable review of the Canon 7D thread" two days ago?