I've been curious lately how light passes through lenses, is there a cutaway diagram of a larger lens like a 400mm that shows how the light bends and turns upside down inside the lens before it hits the imaging sensor?
That's an excellent article, though it doesn't directly address the OP's question.
You might try searching "ray trace diagram" and adding some qualifiers. I've seen samples in old lens tests, but didn't turn up anything for a long telephoto in a quick look.
Yea I was actually looking for diagrams for how light passes through a lens and how it bends as
it passes through each element. There's got to be some out there.
I doubt Canon would release these. And I also doubt anyone would be able to reproduce something such as this, unless they're willing to dissect a lens and do a TON of work. You'd have to go to Canon to get a bunch of details about lens construction that I'm betting they wouldn't disclose.
Also - the light doesn't turn upside down, the sensor records a virtual image.
meechahel wrote:
I doubt Canon would release these. And I also doubt anyone would be able to reproduce something such as this, unless they're willing to dissect a lens and do a TON of work. You'd have to go to Canon to get a bunch of details about lens construction that I'm betting they wouldn't disclose.
Also - the light doesn't turn upside down, the sensor records a virtual image.
Interesting.. I thought that when it hit the sensor that it was upside down.
Anyway, why do you think Canon wouldn't release these? Proprietary lens design that could help other companies?
For instance, the 400 2.8 I'm getting tomorrow, I'm just wondering how it gathers the light and how it acts as it passes through the lens groups before it hits the sensor.
From my take, it's Canon's engineering efforts and designs. I don't think another company could reverse engineer it further, but it just doesn't strike me as something Canon would make publicly available. A lens like the 400 2.8 is very difficult to design and takes a lot of trial and error to find out what works right. You could really do it with just one or two elements, but the aberrations and other artifacts would be downright horrible.
Also yes, in it can be thought of as upside down...but it's referred to as a virtual image. Not all optical systems work this way though.
meechahel wrote:
I doubt Canon would release these. And I also doubt anyone would be able to reproduce something such as this, unless they're willing to dissect a lens and do a TON of work. You'd have to go to Canon to get a bunch of details about lens construction that I'm betting they wouldn't disclose.
Also - the light doesn't turn upside down, the sensor records a virtual image.
it's not a big secret in the general sense you can just ray trace through their diagrams, although they are so complex it would be a major PITA
as for trade secrets they already publish the diagrams, of course they are not manufacturing level detailed and you could always slice a lens in half and carefully measure each element and test the composition
of course you still need to be able to manufacture precisely and make each type of glass element and carve and polish them etc.
meechahel wrote:
Also - the light doesn't turn upside down, the sensor records a virtual image.
A mirror behind the lens reflects the incoming image to a corrective prism. The prism turns the image upside-down, since a lens always reverses the orientation of a picture, so the photographer perceives it as upright.
Imagemaster wrote:
A mirror behind the lens reflects the incoming image to a corrective prism. The prism turns the image upside-down, since a lens always reverses the orientation of a picture, so the photographer perceives it as upright.
Right, but the sensor still sees a virtual image as soon as the mirror flips up.
jorkata wrote:
Most (if not all?) lens designs are patented, so you can find the diagrams in the patent office. They are publicly available.
I stand corrected, I had no idea! These are pretty awesome. And the thought of a 14-24L is even better
The image captured on a sensor (film, CMOS, CCD, etc) is a "real" image, not a "virtual" image, at least as far as those terms are conventionally defined in optics.
meechahel wrote:
Right, but the sensor still sees a virtual image as soon as the mirror flips up.
Yes, I know that. As far as I know, every Canon lens inverts the image. Stick any Canon lens on the front of a view camera and the image on the ground glass will be upside-down.
Imagemaster wrote:
Yes, I know that. As far as I know, every Canon lens inverts the image. Stick any Canon lens on the front of a view camera and the image on the ground glass will be upside-down.
if any camera lens made the image right side up you'd need some sort of magical sensor outside of the camera in front of the lens to grab it
skibum5 wrote:
if any camera lens made the image right side up you'd need some sort of magical sensor outside of the camera in front of the lens to grab it
