As an engineer in manufacturing, I have to say that there manufacturers cannot always or rarely produce a "perfect" product. All copies that come out of the manufacturing line is built to a certain tolerance. So the aperture of F5.6 +/- 0.33 f-stop may be acceptable to the manufacturer. So what if the 17-85 may be 1/3 brighter at F5.6 than the other lens or vice versa? To the manufacturer is acceptable and within it's manufacturing tolerance. Don't lose sleep over it.
The 24-105L, being a constant aperture and a L lens, would have more elements then the 17-85 zoom. More elements = more light being refracted. Thus, a f/4 isn't standard on different lenses, but very close to. The ancient Canon 50mm f/0.95 is one good example, the "real" aperture is closer to f/1.2.
f/8 and be there.
ok thanks everyone for remembering the name :P.
metal, glasses, plastic and a eye to see
I support your initiation in this subject.
Some of people just dont understand where you coming from and insist why bother.
From an engineering point of view, it is indeed a little puzzling.
The next question is if you take both shots, one lens at 1.8 and the other at 1/10. Would they be similarly exposed?
Just wonder did you go into Astronomy?? ha ha ha
Anyway, theoretically when light enter from air to an uncoated lens, it can loose over 4% of light transmission in reflection and scattering per surface (one lens got 2 surface). This light loss will exponentially multiplied when adding more lenses. For example, a single uncoated lens will have light transmission of 92.16% (0.96x0.96) and 2 uncoated lens will have light transmission of 84.93% (0.96x0.96x0.96x0.96). This is the reason why you can see your reflection on glass windows if you are on the brighter side.
A single layer of anti-reflection coating can reduce the loss to 1.5% and a multilayers composed of different coatings (multi-coated) to as low as 0.25%.
So EF-S 17-85mm f/4.0-5.6 IS USM kit lens have 17 elements(34 surface) and EF 24-105mm f/4.0L IS USM lens have 18 elements(36 surface). I assume both use the same fully multi-coated (coating on all surface of lens) standard (assume of light transmission 99.75% per surface), so the light transimission of EF-S 17-85mm f/4.0-5.6 IS USM kit lens is 91.84%(0.9975^34) and EF 24-105mm f/4.0L IS USM lens is 91.38%(0.9975^36).
Base on the above assumption and theoretical calculation, the EF-S 17-85mm f/4.0-5.6 IS USM kit lens should have a very, very, very small advantage in light throughput over EF 24-105mm f/4.0L IS USM lens if both use the same coating standard.
BUT do you think canon will use the same expensive coating on both their 'L' lenses and normal lenses?? I doubt so. So if you ask me, I think the 'L' lenses should have the better coating over the normal lenses and should have better light transmission over the normal lenses. Then you might ask why you get such result?? I'll assume your test is done properly and consistently. One reason I can think of is the physical design and construction of both lenses:
The EF 24-105mm f/4.0L IS USM lens should have better light baffle than the EF-S 17-85mm f/4.0-5.6 IS USM kit lens. So your result might be due to more stray light entering the EF-S 17-85mm f/4.0-5.6 IS USM kit lens which "fool" the DSLR meter and give a faster shutter speed than the EF 24-105mm f/4.0L IS USM lens.
Anyway, the different of 1/8 and 1/10 is very small. There are more importance factor other than light-tranmission, such as optical quality of the lenses.
Just my S$0.02...
Have a nice day.
Last edited by weixing; 25th February 2008 at 05:39 PM.
hmmm, just thought of something... interesting read btw.
with both pictures taken with the difference lens, could you measure the brightness on both images if they are of same brighness so you know the 1/8 and 1/10 shutter spd is also accurate.
TS let us know ya ?
Interesting topic and discussion.
Btw off-topic a little, I've just made some comparison yesterday between 16-35 f2.8L II and 17-40 f4L (purchasing reason) and the 16-35 consistently exposed 1/3 brighter than the 17-40 in Av mode. From every focal length e.g. 17, 20, 24 & etc and aperture f4 to f16. The environment is showroom lightings, camera on tripod, MLU and cable release.
I'm just a hobbyist and not an engineer or technical expert, but found some similarities in terms of 1/3 exposure differences in this discussion. It doesn't bother me as I learned something new.
Happy shooting always
A good thread
Worth my 5 minutes of reading...
just try take some shots at manual settings (same f-stop, shutter speed, FL...) and compare the photos.
if no exposure difference, what the heck. (err... in second thought, might see some difference? influenced by resolution, saturation, resolving power.... difference between lens design?)
But very unlikely to have significant difference (in this case, between 1/8 and 1/10 s) if talking about handholdability difference or stopping subject movement different...
Last edited by Ljung; 26th February 2008 at 04:43 PM.
From my opinion, there's no significant difference between 1/8s & 1/10s. I dont know how "controlled" actually is your indoor environment but if you like to analyse things like this, you better be damm specific.
I would suggest you to repeat your metering act with only ONE of the lense to see how repeatable is the result. I wont be surprise to see slight changes in the result.
Anyone who wants to dump their L because of low light transmittion, kindly PM me.
Interesting doubt here I dont have much input here, but one question though, are both of the lenses have any filters on? Just curious, if any of you missed out this point...
I think Bro Fragnatic do hit the nail on the head with his reply. FOV would be more than enough to explain the slight different in the EV measure, if the EV is not taken with spot metering.
The FOV will vary from one zoom to other, even if they are both at 50mm. As TS do not have the photo for comparsion, we can all debate till the cow come home.