Canon lenses vs 3rd parties...

[Marsu42]

But by that logic, everything can be fixed in post. Why have good metering, just adjust exposure. Why have a fast lens, just add OOF blur later. Ok, I exaggerate.

Indeed: postprocessing blur certainly is a difference to real bokeh, even using the latest plugins. And metering doesn't matter that much anymore to me when shooting raw with the added dynamic range. Color rendition is a problem if you don't remember how the original situation was, and btw nobody else will either. CAs are really the smallest problem to me since they only occur on a f4-5.6 lens at extremely high contrast borders. Thus, this is what I find important about a lens' iq: sharpness > bokeh >> color >>> CAs

 

[RunAndGun]

I'd rather have an optically superior lens than an inferior one. And if having a fluorite element allows them to remove an element, that allows the lens to be lighter and possibly smaller. Call me old school, and maybe working in TV has something to do with it, too, but I don't want to rely on post. Do it right(or as close to it as possible) to begin with. It's great that we can fix SO much in post nowadays, but I'd rather have as clean an image as possible from the start.

 

[Alangeli]

.. I've got the fluorite-less 70-300L ...

The 70-300L does have fluorite elements.

 

[Marsu42]

The 70-300L does have fluorite elements.

Um, link? Usually Dr. Neuro isn't wrong on something like this...

I'd rather have an optically superior lens than an inferior one. And if having a fluorite element allows them to remove an element, that allows the lens to be lighter and possibly smaller.

The question is: Did Canon leave out the fluorite element out of the 70-300L because they're just cheap and wanted to save the money, or did their research show that CAs aren't a problem with this lens anyway? Well, we'll never know - but if they saved $200 for leaving out something I wouldn't have noticed anyway, me and my purse say it's ok.

 

[Loswr]

Thus, this is what I find important about a lens' iq: sharpness > bokeh >> color >>> CAs

Depends on the lens and the situation. For example, I'd be quite concerned about CA in a shot like this (from a 50L):

Think that one click in LR4 would fix that?

The 70-300L does have fluorite elements.

Really? Can you provide a link that supports that? The description states, "It features two Ultra Low Dispersion (UD) elements for improved image quality and reduced chromatic aberration," nothing about fluorite, and the block diagram of the lens in Canon's technical hall does not show any fluorite elements, either... ??

 

[Neeneko]

In general what you encounter with lenses is the law of diminishing returns, you increasingly pay significantly more for smaller and smaller gains. L glass tends to be significantly more expensive then non-L with fairly marginal improvements, but is fairly consistent in quality.

Non L Canon and Sigma stuff has to be looked at on a lens by lens basis, with some being 'meh' and some being very good. Zeiss tends to be like the L stuff, marginally better with significantly higher costs, but better resale value. There is also, as others have said, compatibility issues since 3rd party lenses are reverse engineered, but this is rarely actually much of a problem since Sigma and Zeiss tend to know what they are doing.

Personally I have some L, some regular, some Sigma, and some Nikkor M39 stuff that I use on my camera and have found that if you get a good lens they are all pretty similar in quality.. provided you avoid the lemons.

 

[Marsu42]

inge.jpg[/img]
Think that one click in LR4 would fix that?

Of course you're right, I was talking of moderate CAs on tele zooms... but looking at your sample (I guess it's from a corner at wide open aperture): Did you actually try to fix this with LR4's lens profile CA reduction? I'm asking because it really does a good job, esp. since there's nothing else purple in your shot - but of course sharpness will suffer.

 

[Alangeli]

The 70-300L does have fluorite elements.

Um, link? Usually Dr. Neuro isn't wrong on something like this...

Sorry, I was wrong, I always thought that this lens also got a fluorite element.

 

[Loswr]

inge.jpg[/img]
Think that one click in LR4 would fix that?

Of course you're right, I was talking of moderate CAs on tele zooms... but looking at your sample (I guess it's from a corner at wide open aperture): Did you actually try to fix this with LR4's lens profile CA reduction? I'm asking because it really does a good job, esp. since there's nothing else purple in your shot - but of course sharpness will suffer.

Not my shot, it's from photozone's review of the 50L on APS-C, and I'd guess it's the full image, not a corner crop (this is longitudinal CA, which affects the whole frame, unlike lateral CA which is worse at the edges).

 

[AJ]

Yes you should consider third-party lenses.

For years I've enjoyed lenses from Tamron, Sigma, and Tokina. Zeiss is another good brand. All these manufacturers make superb lenses as well as duds. Do your research before buying one of these lenses to know what you're getting. If you don't want to do your research then buying a Canon L is pretty much a guarantee of quality.

As for sample variance. All manufacturers including Canon, Nikon, Tamron and Sigma make duds. It's probably true that Tamron and Sigma are more prone to manufacturing defects like decentering. That said, buying an L lens is no guarantee. My friend bought a badly decentered 70-200/4. Canon eventually fixed it, but it took six months. Read the photozone reviews about 24-70/2.8L and 24-104/4L for more stories of bad copies of L lenses. If you're concerned about getting a bad sample, buy from a bricks+mortar shop who will stand behind their products with exchanges and refunds.

I think it's a real shame - IMO Canon should test every individual 70-300L before putting it on the shelf. This way "L" would actually be a guarantee of quality. "L" would actually mean something.

It is said that 3rd party manufacturers are no insurance for future compatibility. Perhaps this is true. However Canon has changed their lens mount too, rendering old lenses useless. Today, a lot of FD glass sells for pennies to the dollar because of this. When Sigma lenses started misbehaving years ago, Sigma offered free chip upgrades.

So, let me make an analogy. Suppose you ask what's good for dinner and someone suggests filet mignon (Canon L). It's certainly a good suggestion, one you can't go wrong with, but there are other cuts of beef out there (Canon non-L) as well as pork and chicken (Tamron and Sigma).

I love my Tamron 17-50/2.8, Tamron 90/2.8, and Sigma 10-20 dearly. These lenses have performed for me year after year. They don't owe me anything, and I don't care about their resale value.

 

[Marsu42]

I think it's a real shame - IMO Canon should test every individual 70-300L before putting it on the shelf. This way "L" would actually be a guarantee of quality. "L" would actually mean something.

Did you have an actual bad experience with this particular lens, or are you citing reviews? I'm asking because I thought a lot about "how to get a good copy" before I bought my 70-300L, and there were very divided opinions about how large the manufacturing tolerance on Canon L lenses really is and how likely it is to end up with a bad copy that needs countless cycles of service and adjustment.

 

[Loswr]

Suppose you ask what's good for dinner and someone suggests filet mignon (Canon L). It's certainly a good suggestion, one you can't go wrong with, but there are other cuts of beef out there (Canon non-L)...

Chuck roast? Please...that's for the dog.

 

[AJ]

I think it's a real shame - IMO Canon should test every individual 70-300L before putting it on the shelf. This way "L" would actually be a guarantee of quality. "L" would actually mean something.

Did you have an actual bad experience with this particular lens, or are you citing reviews? I'm asking because I thought a lot about "how to get a good copy" before I bought my 70-300L, and there were very divided opinions about how large the manufacturing tolerance on Canon L lenses really is and how likely it is to end up with a bad copy that needs countless cycles of service and adjustment.

I should have said: Canon should individually test every L lens.

I don't have a direct experience with a bad copy of 70-300L. I just picked that one because IMO the price is quite steep, and because it came up previously in the thread.

I've only ever purchased one bad lens myself, a 17-55/2.8 which was very blurry at 35 mm somehow. I sent it back and got a better copy.

The faulty 70-200/4L that I referred to was bought by a close friend of mine. In fact we placed a joint order to save on shipping. He'd been shooting shallow-dof photos for a few weeks and chalked up strange results to user error. Then he ran a brick wall test. He showed me the problematic photos. One side was severely blurred even when stopped down the f/11. At this time the return policy was up so he had to deal with Canon.

 

[Loswr]

I should have said: Canon should individually test every L lens.

Boom. The cost of all L-series lenses just went up by 15%. Any manufacturing process produces a few lemons, and QC is designed to reduce that number to an acceptable level. 100% perfection and/or testing every lens simply costs more than it's worth. That's why there's a warranty.

 

[dstppy]

I should have said: Canon should individually test every L lens.

Boom. The cost of all L-series lenses just went up by 15%. Any manufacturing process produces a few lemons, and QC is designed to reduce that number to an acceptable level. 100% perfection and/or testing every lens simply costs more than it's worth. That's why there's a warranty.

Maybe THAT'S the deal with Canon's latest pricing

I'm honestly curious how this new Tamron 24-70 f/2.8 VC is going to stack up to the outgoing EF 24-70mm f/2.8L USM . . .

 

[bornshooter]

I should have said: Canon should individually test every L lens.

Boom. The cost of all L-series lenses just went up by 15%. Any manufacturing process produces a few lemons, and QC is designed to reduce that number to an acceptable level. 100% perfection and/or testing every lens simply costs more than it's worth. That's why there's a warranty.

100% perfection is impossible in all thing's in life..

 

[AJ]

Lens testing the way Canon does it is actually a pretty quick procedure.

The lens is used much like a slide projector lens: a test pattern is back-projected through the lens from the lens mount out the front and onto a screen. I imagine a technician could can test a prime in less than a minute or so, a zoom in a minute or two, assuming no adjustments need to be made.

 

[lol]

For relatively high cost low volume items like DSLR lenses I'd be very surprised if they don't already test 100% of lenses made. However, they might not test them all the same way i.e. they don't 100% test 100% of lenses. For example, they would at the least give it a function test to look for any obvious flaws and a basic optical sanity check. As long as the testing is optimised for the manufacturing process the cost impact is minimal. If you demanded the head lens designer to individually inspect every lens, that would be a different story.

 

[peederj]

If you like automation features, certainly get Canon lenses with a Canon body. The luxury line is abbreviated "L."

If you have money and work manually, maybe get Zeiss primes. You will find ready buyers for those second hand if you wish to part with them.

If you don't have money and work manually, try Samyang/Rokinon. Or if you are afraid of taking valuables on location but still want an excellent manual capture (e.g. guerrilla video).

If you have a moderate amount of money and want some of the automation features, try Sigma/Tamron/Tokina which might give you a little and take a little vs. the Canon equivalents.

I have examples of all of these...some are better deals than others, but none that I chose were ripoffs.

Have fun and take pictures.