Comparison test for vignetting: EOS 3 vs 5D vs 1Ds II

More...

More yet...

and yet more...

Now for the controversial one - the 24-105mm f4 L....

Lastly...

Thanks a lot for the test, James! It is always better to base a discussion on facts instead of opinions.

This makes it clear that all full frame cameras - digital or film - vignettes with a wide angle lens set at maximum aperture.

I think much of the confusion about full frame cameras and vignetting comes from the fact that most people shot with 1,6 crop cameras - me included. And we are not used to see pictures with this much vignetting. So we think, that there must be something wrong - either with the lens or with the camera.

But since we use the same lenses on our cropped cameras without any severe vignetting, we think that it must be the cameras fault.

Anyway these pictures makes me glad, that I recently bought a 30D instead of a 5D. Much less postprocessing to do in Photoshop.

"I think much of the confusion about full frame cameras and vignetting comes from the fact that most people shot with 1,6 crop cameras - me included."

True. However, it was only 5 years ago most of us were shooting with full frame film. I find it amazingly so many were so inobservant. Light fall-off was always there and more so with wide zooms.

I have only a few slides with light fall-off from the last couple decades. I quickly learned not to shoot skies with fast apertures when using wide angle lenses. I find it odd so many 5D owners will blast off hundreds of shots at 24mm wide open and then cry foul.

Nice job James!

I find it odd so many 5D owners will blast off hundreds of shots at 24mm wide open and then cry foul.

With all due respect, it's not the 5D owners that are "crying foul." It's trolls like Steven Moseley who don't even own Canon equipment, but just like to be controversial and - as he admits - likes to "ruffle EOS owners".

I guess some people just need to get a life!

I applaud your bravery! In the face of those who say "the 5D vignettes, even Canon UK says so!". LOL.

Well done; definitive; story over. Case closed. Nicely posted too.

-Ken

"Case closed."

Hardly. The dull and ignorant will never understand nor be satisfied with such clear evidence. They will continue to rage unchecked forevermore.

Thanks James - that seals he point I made in my post "The Canon 5D does NOT vigentte"

;p

5D-owners are funny. They look at pictures taken with their camera, which shows horrible vignetting. But they celebrate that the vignetting in the pictures taken with another camera is just as bad.

The 5D doesn't vignette. But pretty much all the lenses you put on it does. That's just great!

However, it was only 5 years ago most of us were shooting with full frame film. I find it amazingly so many were so inobservant.

Yeah, but if they shot negative film, the prints typically excluded the edges and corners, and if they shot slide film, the slide mount often cropped off a bit of the edges and corners. So the very worst areas were not visible in the final image. That doesn't explain the whole thing but it may explain why people seem to notice it so much more now.

thanks for the test, nicely done.

>>But pretty much all the lenses you put on it does<<

Find us a 24mm that doesn['t show light fall at f/1.4 and we'll be happy to use it.

Until then...

Or...the 5D and 1Ds II and EOS 3 all cause it :)

Sorry couldn't help myself!

m

>>Or...the 5D and 1Ds II and EOS 3 all cause it :)<<

Trust me, some people actually believe that!

Like the dude that posted above...they still can't *grasp* that light fall off is a lens characteristic that will show on ANY camera, ANY format, ANY brand with ANY lens.

Aleluya! (sp?)

"The 5D doesn't vignette. But pretty much all the lenses you put on it does. That's just great!" Well, lenses have always "caused" light fall-off in images, nothing new there.

Great test, James. As a former film user, I expected similar results and your test shows just that.

The 35mm f1.4 (@ f4.0) looks to win by a wide margin. Be curious to see how the 24-70 fares. It quite well here:

http://www.luminous-landscape.com/reviews/lenses/28-105.shtml

Just a thought: the IDSII vignetting should be *slightly* more pronounced than the 5D, all factors being equal: it's sensor is slightly bigger.

the IDSII vignetting should be *slightly* more pronounced than the 5D, all factors being equal: it's sensor is slightly bigger.

The 1Ds sensor is 0.2mm longer and 0.1mm taller than the 5D--about half of 1 percent bigger.

That sounds about right. I did say *slightly*. Wouldn't the light fall-off increases logarithmically, making the diff more signif than .5%

No surprise here. It is actually the first time that I see results from the 24-105 with slide film. When Canon released the 24-105, I thought: good, this is the lesn I have been expecting for my travel photos. Then I saw the Chasseur D'Image test and I was just amazed. So much light fall off, even at f/8. Clearly, this lens would be unusable for me. I do a lot of photography from mountain hikes and in the desert, involving lots of blue skies, so...

I think lens designers are cutting corners in their work, knowing that these defects are easy to adjust while post-processing. Well, if a lens costs more than 1000 USD, it ought not to show such a poor performance regarding light fall off.

I don't quite understand what all the complaining is about. If I'm shooting landscapes, I would probably be using the wide end of my 24-105, but with the aperture stopped way down to get maximum depth of field, in which case the light fall-off would never be an issue. If I want to take portraits, I would be using the tele end of the 24-105 at f/4 to get background blur, but light fall-off is demonstrably not an issue when the lens is extended.

For those rare instances where I want or need a wide field of vision shot at the widest apertures, I could compensate by framing the image large enough to allow later cropping of the "undesirable" areas. Does that make this lens a dog? No more than if I had to exercise care to avoid lens flare or distorted images of buildings, for example. It seems the magic of digital photography has caused some to expect the physically impossible from their "professional" caliber equipment.

Although the 24-105mm - at least my copy - does still exhibit a bit of vignetting at f8 it is totally gone at f11. But even at f8 it really isn't much of an issue at all.

I just came back from a month in Polynesia, Easter Island and the Atacama desert of Chile and the only lens I had for my 5D was the 24-105mm (the 'serious' photography was being done in b&w 6X6). When I was handholding the 5D it was almost always at f8 with IS on and there isn't one shot where you would notice vignetting - if you went looking for it of course it was there but it was never at a disturbing level. I have shown that portfolio of shots to over a dozen people now including a couple of people who would notice that sort of thing. Not a peep out of them on that score - just compliments on the amazing quality and detail of the shots.

Paulo, believe me the 5D is an excellent hiking/landscape lens and I unhesitatingly recommend it. Also, when you are in harsh bright desert/mountain light you can easily handhold it at f16 and get pin sharp images because of the IS. Shoot at ISO 200 too as it is noise-free on the 5D. Once you know where the vignetting starts and you keep it level to tame its distortion (only noticeable for water shots) then it produces brilliant results.

First of all, I 've already posted a reply about this subject, maybe in a different thread but it seems desappeared. By the way, many canon 24-104 in combination with 5d do actually present a dramatic vignetting - please let's not start an argument about the meaning of vignetting... let's just take it as it is: a darkening of corners due to SOME reasons - and this is CONFIRMED BY CANON ITSELF. some 24-104 lenses recognizable by serials are flawed! This means that 5d and 24-104 owners who complain for too much vignetting are not necessarily ignorant. some of them are true!

Anyway it seems to me that even flawless 24-104 vignette more than many other L lenses.

The 5D does not vignette; the 5D is as good a camera as there is whether digital or film; it is the 24-105mm lens that vignettes amongst many many other examples. This is what this whole test was about. It is the lens alone that causes vignetting, or if not alone then is the main cause of vignetting, and there is no debate that the 24-105mm can produce quite a lot of it. But there is nothing at all wrong with the 5D itself as a camera body that I can see from my test.

All my test proves is something that most people knew already, lenses shot wide open vignette and fast wide-angle lenses shot wide open vignette a great deal. That has always been the case across all lenses on any brand of camera. I am surprised that so many people think they have uncovered some nasty lurking trait of the 5D when it is demonstrably the case it is behaving normally and as expected.

I have not run any controlled tests, but having both the 3 and the 5D, I've found that my lenses behave the same on both bodies. Same degree of light fall off. Sometimes perhaps more noticeable with the 5D, perhaps because film grain on my scans tends to mask the fall off a bit. Density seems the same though.

Another non-issue except for the people who always see problems where none exist.

Hello again..

Firstly..an excellent test that shows that the 5D owned by James is acceptable.

Secondly..to John Bellenis: I have no problem with critisiscm, but please get your facts correct and do not mis-quote me...I did NOT say I liked ruffling EOS users...so please re-read my posts, as I did not say that. I just observed that it was easy to do so...which is not at all the same thing.

The original post I placed was intended to warn potential buyers that there may be better options than a 5D. Also John... it IS 5D owners who ARE crying foul..there are some on this thread and also several other 5D owners on other forums also crying foul. If you think 100% of 5D owners are happy, you are not reading what is put before you, or just do not want to accept it. If you are happy, then fine, but please, do not be so arrogant as to tell other 5D owners they should be happy, or that others do not have the right to publicly express their views or issues. We are not in some socialist state here.

I am constantly stunned that when someone has the 'temerity' to express a view, then almost immediately, several owners of the said equipment take this 'view' as some personal assault on their own equipment..what happened to free speech? what happened to reasoned thought?

On this thread I see at least two other people who have the view that the 5D has issues..you gonna stone them also?

Cheers Steve.

It'd be nice to also try subjects that are up close.

JAMES: I agree with you but you've well demonstrated that a non flawed 24-104 behas quite the same way on each ff camera you used for the test. I don't say d5 is a bad one, and i'm sure must of the problems depend on 24-104 but we couldn't see any picture taken by a [b]flawed[/b] 24-104 and d5. it is possible that flawed 24-104 have some mechanical or optical difect that is more evident - maybe for design reasons I actually don't know - on d5 for is digital sensor dimension. we know digital sensor are different one from another even if both full framed and are more sensible than films to the angle by which the light falls onto them; this is why some lenses like 24 2.8 are good with analogic cameras and produce poorer images on digital ones. this of course doesn't mean it's a d5 fault completely or d5 is bad designed.

>>many canon 24-104 in combination with 5d do actually present a dramatic vignetting... and this is CONFIRMED BY CANON ITSELF. some 24-104 lenses recognizable by serials are flawed! This means that 5d and 24-104 owners who complain for too much vignetting are not necessarily ignorant. some of them are true!

The earlier recall was for an issue unrelated to light fall-off. Canon has never stated that there is anything abnormal or defective about this lens' propensity to "vignette".

>>Anyway it seems to me that even flawless 24-104 vignette more than many other L lenses.

Then you should take another look at James' test results. You will find that images of other wide-angle L lenses used at wide open apertures show the same degree of light fall-off on all 3 cameras.

On this thread I see at least two other people who have the view that the 5D has issues..you gonna stone them also?

Giving people sh*t because they make unsubstantiated claims is hardly the same as stoning them. Why the ridiculous melodrama?

[And by "unsubstantiated claims," I mean that I have never seen anyone who was complaining about their faulty camera show any pictures demonstrating that they had more fall off with the 5D than with a film camera.]

Guido,

I know what you are saying but somehow we have to separate the 24-105 and the 5D in order to understand clearly what the problem is. I know some people have been complaining about the combination of the two but what would conclusively prove that the 5D on its own has a problem is if someone could produce a photograph of the sky taken at f22 with dark corners. Then we have some juicy evidence!

I am pretty sure that some users were surprised at the vignetting they got from their new 5D/24-105mm kits and have blamed the camera instead of blaming the lens - particularly if they have upgraded from a 20D or the like where the nasty dark bits were cut off. Full frame digital mercilessly exposes the faults of lenses and I firmly believe that that is all that has happened here.

Say hello to Milan for me - I lived on Corso Italia for four years.

Ciao,

James

>>test that shows that the 5D owned by James is acceptable<<

You still haven't answered my questions from my other reply to YOUR posts saying "the 5D is flawed". The questons are simple:

1) Provide some basis and data to support your asserting that the 5D is a flawed camera

2) is YOUR 5D also flawed?

and I will add this new queston:

3) can you post some samples of your 5D images showing the "flaw" of which you speak?

Those a reasonable questions that deserve an answer, especially if you want to be taken seriously.

JAMES: ok i think I got you. I didn't know so many people thought the problem were completely due to 5d. I'll say hallo to milan from you, thank you! Corso Italia is 10mins by walk from my house. I'm curious about your the time you spent here.

IORI: didn't canon admit there were some vignetting problems with 24-104 lenses with serials starting with UT, at almost every apertures? what did they say exaclty - sorry maybe i had some wrong information -

by the way I said it's eveident the fall off compared to others L lenses as the shots above were shot at f1.4!

In short, those who shutterbugs before the era of digital camera knows that lens vignets at wide-open, nopt just wide-angles but lenses longer focal length. The effect may be alleivated to certain extent by more technically advanced design (the more expensive). You can also completely eliminate vignetting if you adapt a medium format lens to use on a 35mm body.

It's a matter of physics the the "usable" image coverage by lens designed for a particular format may not be large enough to cover the full frame. Thus a 1.6 cropped sensor is a way to solve this problem.

Guido: The recall was for a lens flare issue. See: (link)

Hi Mike,

I do not have any melodrama, some of the 5D owners on here have made one, which is why i was asking why on earth these folk take these things personally? it seems rather silly and juvenile to say the least.

Also, you seem to think it is OK to give people s*** as you so politely say...WHY? do you not believe in free speech either? and how about politeness?

And...in the first instance, i never made any claims, either unsubstantiated or otherwise, I just directed people to a thread on another forum where OTHER people; in fact some 5D owners who were voicing their OWN unhappiness with the 5D...what is phot.net users problem with this? it is a lot of them making the 'melodrama' as you put it...and their claims are not unsubstantiated either..there are clear photos showing the issues very clearly...and it is far worse than any light fall-off i have seen with any other camera/lens combination.

Another user on here then posted some so-called proof photos which claimed to show the problem is non-existant..only the photos DID show the issue and the poster conveniently could not see it himself...strange..also they were cropped which the poster somehow forgot to mention...

One of the few posters on here being polite and sensible is James, the OP of this thread.

I remain shocked at the confrontational attitudes on here, which is just silly.

Cheers Steve.

LOL - Wow Steven - a little thin skinned or what?!

Take a chill pill - it's just a little lens issue for a system you don't even own! As to being misquoted... ahhh virtually evrything you just ascribed to me, I never said! Oh well, try to keep the old blood pressure down, it's really not that important!

Steve: I think we're all just interested in the truth; not just speculation by those who are unhappy with their new 5D/24-105 combination because they can't accept a certain amount of light fall-off in their equipment. You state that there are images posted by owners that show light fall-off that is much worse than anything shown by James' images. Can you please point me to a link, as I would be interested in examining them. Thanks.

>>in the first instance, i never made any claims, either unsubstantiated or otherwise,<<

So Steve, do you actually own a Canon EOS 5D?

You DID make a lot of claims and since you haven't posted ANY pictures taken with YOUR Canon 5D I must assume you don't have one!

Here's some of the claims you have made:

"The 5D does vignette"

"posters on here do not want to hear that their beloved Canon DSLR is flawed"

"the 5D DOES have issues"

"many buyers have bad 5D's"

So, I guess since you don't call those statements claims we should just put them in the BS bin :)

I think I'll stick to the 1.3X for now..... :-)

Happy shooting, Yakim.

"The original post I placed was intended to warn potential buyers that there may be better options than a 5D."

What is the source of your 5D expertise? Why should anyone heed your advice?

Good effort James, well done.

I believe this whole thing started by someobe, MISreading the note from Canon. maybe the next model will have LFO elimination built in the body software.

Hi Iori,

The Link is at the bottom. It is a long thread, but stick with it, as it is illuminating. The photos are about half way I think. This is the same thread that users here did not like much..but take your own views. There are follow up theads also on the same forum. The real difficulty is that different 5D owners are having different amounts of corner darkening..so either some owners are not seeing things too well, or different examples of the 5D are performing differently, which would indicate a Canon quality control issue..maybe with the sensors?

I agree with you fully about wanting the truth..but the difficulty seems to be that some people quite literally 'see' things others cannot and vice versa. So the 'truth' is very hard to decide. The issues, by the way are with a variety of lenses, not just the 24-105.

(link)

Cheers Steve

Hi Giampi,

Take a look at the link I posted for Iori and the other threads on this issue on the same forum and read and look carefully...then make up your own mind...

Cheers Steve.

John...excellent advice, I have just received a chinese herbal potion from my accupuncturist, so I hope it may help the chilling process....

Gary...no one need heed any advice from anyone on here, it's just a public forum to post views...check out the link on my previous post and make up your own mind...

Cheers Steve.

Can you try on Nikon 17-35mm f/2.8 on it? or zeiss 21mm?

Thank you for the test, may I conclude, that all 3 canon full frame, combined with 3 lens 24,35, 24-105 lens are all vignetts. This is a very very good proof and facts.

The only bad thing is, I found this fact AFTER I bought my Eos 5D and 24-105 IS expensive lens.

Now I understand the flaws is in the lens.

I deeply regret, I own 5D perfect body, but cannot paired with 24-105 lens, 24, and 35 because it vignett in large aperture. No subtitute lens able to overcome this problem. And my conclusion will be: Buying EOS 5D is my biggest mistake (I didnt say 5D has a defect/problem)

Any one who wants to call me crying foul please wellcome, I am a foul, bought this camera, and i am crying because i spent US$ 3,500 and found these facts. I dont care of 1DS Mark II or EOS 3 has same problem with these lens, because I have no more chance to own them, I spent all my saving to buy 5D and 24-105 IS, the best match from the best brand as the dealer advertise.

why not just learn to avoid it (don't use wide open at 24mm, stop down when you can), and learn to fix it (trivial in ACR and other RAW converters)

In consumer electronics, in general, the more money you spent the better equipment you get, and most important, better in almost every single respect. Better sound, better image, better reception, etc. In the photographic world more money doesn�t necessary means better pictures and I think that a lot of the new people taking photography come with this mentality and they just spend as much money they can without fully understanding the equipment they are buying and they end up feeling cheated buy the manufacturer.

Regards

Rafael

Errata:

"by the manufaturer"

Rafael

"Thank you for the test, may I conclude, that all 3 canon full frame, combined with 3 lens 24,35, 24-105 lens are all vignetts. This is a very very good proof and facts.

The only bad thing is, I found this fact AFTER I bought my Eos 5D and 24-105 IS expensive lens. "

Sigh...1.6 crop, what damage hath thou wrought?

I guess some people will never get it.

Regards,

Bishop

Those are three good lenses to choose if you want to "prove" there isn't a vignetting problem. Fast wides typically have exit pupils about 80mm from the image plane, and medium zooms close to 100mm.

You could just as easily pick a 20mm f2.8, 35mm f2, and 17-40mm zoom and "prove" that there is a vignetting problem.

;)

I thought vignetting is a property of a lens not a camera? Silly me.

Hi Mario..

actually this light fall of issue is a result of a combination of lens WITH a specific camera...

Thank you for the response.

First thing first, I used Film Camera Minolta 500xi, with 20mm Tokina Lens since 1996. I never saw light fall off or Vignett.

In 2005 September, I bought Digital KonicaMinolta Dynax 5D (with crop factor, surely I never found such a thing (I used Tamron lens 11-18mm, with Hoya CPL lens filter attached))

In April 2006, I upgrade to EOS 5D, with 24-105 IS lens, and here I am in this forum.

Quote from Rafael: "In consumer electronics, in general, the more money you spent the better equipment you get, and most important, better in almost every single respect. Better sound, better image, better reception, etc. "

I think this is a very common way of thinking most of customer, including Canon SLR EOS majority consumer.

quote continues..

In the photographic world more money doesn?t necessary means better pictures and I think that a lot of the new people taking photography come with this mentality and they just spend as much money they can without fully understanding the equipment they are buying and they end up feeling cheated by the manufacturer.

This is really new for me.

My friend is Nikonian, he upgrade from d70s to D200 then to D2X. The more money he spent, the better the goods, the better image quality he gets.

I thought this also applied to Canon, but seems im wrong. In Canon, more money spent, requires understanding (and accept)law of physics, and increase ability to work around the sophisticated equipment, also require PhotoShop as a compulsary Software.

Quote from mr Bishop.

I guess some people will never get it.

Thats correct sir, I just dont get it. How come my 1997 full frame film Camera with 20mm lens did not vignet, and the combination of 2005 EOS 5D Digital full frame Camera with 24-105mm has light fall off and vignett.

Canon should put some kind of warning in the advertise or on the box said: Warning, Full frame sensor, Light fall off might occured with following lens kit: 1., 2., 3., etc.

Regards.

I 100% agree with Steven.

How many people ever worked wide open with a wideangle shooting landscapes on film? Not many, I would suggest. So they never saw the vignetting. With digital, people try things out because they can at no cost, even if it bears no relationship to their usage. And then they're surprised at the results.

"actually this light fall of issue is a result of a combination of lens WITH a specific camera..."

I don't get it, apparently you didn't read the thread?

���How come my 1997 full frame film Camera with 20mm lens did not vignet,���

I only shoot film and at one point I owned an Elan II (eos 3 now) and the same 20mm f2.8 canon lens, and LFO in one of the reason I sold that lens almost 6 year ago. Did you used to shot negative or positive film?

Regards

Rafael

You need to fix the other parameters of the imaging system. Also the sky is not a good subject for this test as its brightness may vary depending on angle. You need a homogeneous surface with diffuse lighting to test this.

Now taking a few quick readings of the 1Ds MK II image and the EBX image in the test, there seems to be some darkening of the sky in the DSLR shot midway between center of the image and corner, which is not exhibited in the film image. This same phenomenon does not show in the buildings in the center of the image, hence it seems that the FF imager increases contrast due to vignetting vs. actual subject contrast.

But it's impossible to say for sure in such a test with so many variables. The shots aren't even framed exactly the same. I recommend doing the same test in a diffusely lit room - just photograph a clean wall with a reference gray scale. The reference gray scale is needed to normalise the curve of the imaging system.

Nice to see you here. do you notice that the image sensors of both 1DS II and 5D is not at the relative center of lens/image circle ? Hugo So / UR GALAXY

Rafael - "How come my 1997 full frame film Camera with 20mm lens did not vignet"

Because film isn't particularly sensitive to the exit pupil location of the lens. The common Canon or Nikon 20mm f2.8 has an exit pupil about 52mm from the film plane, and the image circle radius of 35mm film is 21.6mm (sqrt(24^2+36^2)/2). So light strikes the corners of the image at 22 degrees (arctan(21.6mm/52mm)) from perpendicular. That doesn't bother film at all. Heck, film isn't even troubled if you have a 20mm rangefinder lens, with an exit pupil just 20mm from the film plane, and light striking 47 degrees (arctan(21.6mm/20mm) from perpendicular.

But digital is different than film. Due to the microlenses, a typical digital sensor has about a full stop less sensitivity to light 22 degrees from perpendicular than it does to light that actually is perpendicular to the sensor. So you get an extra stop of vignetting in the corners, added to whatever vignetting the lens already has.

Other wide designs have "better" exit pupil locations. A 24mm f1.4 is out more than 80mm from the film plane. So the angle is something like 15 degrees from perpendicular, and the vignetting is less than 1/3 stop.

Joseph Wisniewski: "Due to the microlenses, a typical digital sensor has about a full stop less sensitivity to light 22 degrees from perpendicular than it does to light that actually is perpendicular to the sensor."

There is no such thing as a typical sensor. The technology is very much in progress. If you check specs for the latest generation of Kodak MF sensors (http://www.kodak.com/US/en/dpq/site/SENSORS/name/ISSFullFrameProductFamily , look at KAF-39000 and KAF-31600 datasheets), you will find that they lose less than 1/2 stop at more than 30 angles. There is no reason to think Canon cannot do as well or better.

Aside from the theory, I have checked the sensor vignetting using a 17-35/2.8 at 17mm on slide film and 5D. The difference was less than 1/3 stop of the three stop corner vignetting. I did a fairly scientific test, but I think everyone will argue this one ad infinum, or possibly till they try it themselves with their own lenses for the exact same shot using similar contrast film/digital processing.

Thanks Joseph, but I was quoting Riza.

Regards

Rafael

Hi Hugo! Nice to see you here!

I haven't noticed any difference in the centering of the images but I'll have a look when I get the chance.

See you soon online no doubt!

James

Radek Tezaur - "There is no such thing as a typical sensor."

Why, of course there is. And your own example only proves my point.

"The technology is very much in progress."

That is true. The technology has converged on the point that exactly matches my definition of typical.

"If you check specs for the latest generation of Kodak MF sensors"

Well, I would have, were this discussion about medium format photography.

"look at KAF-39000 and KAF-31600 datasheets), you will find that they lose less than 1/2 stop at more than 30 angles."

Look at them, yourself. As you said, they're medium format. For a number of historical reasons, generally involving the desire to insure the cameras are compatible with shift lenses and small view cameras such as a Arca, medium format sensors typically have no microlenses. They also typically have ISO ranges from 25-200.

35mm full frame and smaller "APS" sized cameras typically do have sensors with microlenses. Users of such cameras are not satisfied with a maximum speed of ISO 200 or 400. I can only recall two examples of cameras with APS or 35mm sized sensors that did not have lenses, and both failed in the market place and are now discontinued. amples I can recall that did not are both out of production. The Sigma SD9 was replaced in 2003 with the SD10, which added microlenses to the sensor used in the SD9. The Kodak SLR/n and SLR/c were discontinued in 2005.

"There is no reason to think Canon cannot do as well or better."

Of course there is. There are multiple reasons. First, Canon is not currently in the medium format market, and there is no reason to believe that they are going to enter that market. Second, Canon is most known for their high ISO performance and most aggressively pursues the sports and photojournalist market. They are not going to compromise those markets to introduce a camera with diminished sensitivity.

Radek Tezaur - "Aside from the theory, I have checked the sensor vignetting using a 17-35/2.8 at 17mm on slide film and 5D. The difference was less than 1/3 stop of the three stop corner vignetting. I did a fairly scientific test..." (elipses mine)

Sorry, but you didn't do a "fairly scientific test". You did pretty much the same test that James Symington did. You tested one lens, and drew your conclusions from too small a sample base.

Read what I wrote farther up this thread, about the lenses James used "Those are three good lenses to choose if you want to "prove" there isn't a vignetting problem. Fast wides typically have exit pupils about 80mm from the image plane, and medium zooms close to 100mm. You could just as easily pick a 20mm f2.8, 35mm f2, and 17-40mm zoom and "prove" that there is a vignetting problem."

Now, I've never measured the exit pupil of the Canon 17-35mm f2.8, but the Canon design is similar to the Nikon 17-35mm f2.8, and I have had that on the optical bench. For the Nikon, the exit pupil is 98.0mm at 17mm. This will cause no problems for a sensor with microlenses. It also tells you nothing whatsoever about how the camera will perform with lenses like a 20mm f2.8 (exit pupil at 53mm) or a 50mm f1.8 (exit pupil at 54mm).

Joseph, is the exit pupil distance you're talking about related to the "telecentricity" mentioned by some makes regarding their wide angle zooms?

Andy, exactly.

They don't typically make lenses that are truly "image side telecentric", with exit pupils infinitely far from the sensor. (There's also an "object side telecentric" lens, with an entrance pupil infinitely far behind the camera. Those lenses are interesting, because they have no perspective).

They do make lenses that are "near telecentric", that is with exit pupils sufficiently far from the sensor to eliminate any angle of incidence problems. I believe it was Olympus, in their four-thirds (tm) literature that first used the term "near telecentric", and defined it as having the chief rays within 8 degrees of perpendicular at the corners of the sensor. That may be a bit much, just about every sensor I've ever worked with is "happy" as long as you're within 15 degrees.

So you're saying the 17-40 is not a telecentric or near-telecentric design? I was impressed with the lack of fall-off on my Nikon 12-24, the 17-40 has some falloff on the 5D but I consider it minor and correctable. Haven't compared it on film though.

Andy, sometimes it's not the exit pupil location. Film or digital, the 17-40mm is simply not one of Canon's finer moments in lens design. Light falloff might simply be from a narrow optical path.

I beleive vignetting is more pronounced on a monitor (light source) than it is on print paper (light relflecting)

Funny...I was looking for the most vignetting lenses available as I just love the effect and read this forum by accident.

Idea nobody had in the discussions above: take off the lens from your camera and shoot without lens....if your camera body has a problem it will show...all the rest is the 'glass'! And vignetting is not evrywhere all the time...otherwise i wasnt looking for the lenses that have it more than others. My 16-35 f2.8 doesnt show vignetting on a FF body.