Another Purely optical adaptor discussion at DVinfo.net

Wayne Morellini · February 22nd, 2007, 08:48 AM

Continuing discussion from another thread:

http://www.dvinfo.net/conf/showpost....05&postcount=9
http://www.dvinfo.net/conf/showpost....0&postcount=10
http://www.dvinfo.net/conf/showpost....7&postcount=12
http://www.dvinfo.net/conf/showpost....2&postcount=13
http://www.dvinfo.net/conf/showpost....5&postcount=14
http://www.dvinfo.net/conf/showpost....6&postcount=15
http://www.dvinfo.net/conf/showpost....9&postcount=16
http://www.dvinfo.net/conf/showpost....8&postcount=17

Quote:

Originally Posted by from latter post

I am sorry I wrote that wrong, I have been quiet sick and tired, ands it has been years since I thought up the scheme. The amount of lens elements in Telecentric is staggering. There is probably some coatings that could handle the deflection, but the idea was to reduce the angel and do it in stages. I should also say, that I am planning to use an sensor with latitude extension, not that I am already. There are some excellent latitude extension sensors coming to market (as far as the latitude they can achieve, but have yet to find out if the picture is nice).

It is something, I had been planing on investigating an few years ago, but got distracted but eh Digital Cinema projects. I am learning a lot, and have some Optical reference stuff I have been planning on reading for the last few years, to brush up. I still haven't got to read that purely optical thread as well.

I now remember, I have an need for telecentric in this and other projects. I want to feed into the telecentric section directly, or to use and macro to resize down to the telecentric section. As an knowledgeable person, could I ask, does this sound practical, or are there problems in what I am thinking?

Would you know how parallel/telecentric (was that the right word) the rays coming out on an 4/3rds lens are?

It occurs to me, if the 4/3rds system is already telecentric, could an 4/3rds lens with macro feature be used as the relay section from an larger lens? I know that I could use an 4/3rds lens directly and relay down, but I like those extra stops.

Ryan Damm · February 22nd, 2007, 04:49 PM

It's actually pretty easy to make a lens system telecentric - all you have to do is put a single element lens in the path with the proper focal length. Essentially, the additional lens should be one focal length away from the exit pupil of the previous lens system. The ray bundles emerging from the second lens will now be parallel.

They will also be severely aberrated. This is why optimized telecentric lenses have 50% to 300% more elements than a non-telecentric lens. In fact, 'doing it in stages' is exactly how most designs handle it: each element contributes some ray bending, and introduces/cancels some aberrations. This is not for the faint of heart; I have a friend who had to design a telecentric zoom lens in college without dedicated lens design software... he said it took a long, long time (sadistic professor, I suppose). Even using software it's daunting - that's why optical engineers are optical engineers.

And could a still digital lens be used as a macro/relay? I don't see why not, but there may be issues I'm not aware of. Your point about the 'amount' of telecentricity is a good one - I believe the lenses are telecentric to avoid acceptance angle issues, which may be more forgiving than the beam splitting prisms on 3-CCD camcorders. If not, the only engineering challenge for you would be securing the proper flange focal length - unfortunately, it may be true that the flange focal length on the 4/3 lens is very short, shorter than the optical path distance from the CCD to the front of the prism block.

You could extend that back focal length with a negative lens between the 4/3 lens and the prism block, but you'd change the FOV (essentially zooming in) and diverge the ray bundles - that is, you'd lose some telecentricity. This would also decrease the effective speed (f-stop) of the relay lens.

And the previous point still holds - the relay lens would have to be awfully fast to get the DOF benefits of the front lens. Again, how fast depends on what your magnification ratio would be. (The magnification ratio is equal to the f-stop ratio; so if your chip is half the size of the target plane, your relay lens must be twice as fast - half the f-number - as the taking lens.)

For my money, if you're going to be starting from scratch with a sensor, I'd look into fiber optic tapers. You'd still have to make the ray bundles somewhat telecentric - but probably a single-element or even achromat field lens would give you enough ray bending to get inside the acceptance cone of the taper. The advantages? No tricky aberrations or optical design software, and you would actually see a several stop gain (tapers are used in night vision for that very reason).

I tried attaching a file (100r.pdf) which talks about tapers a little bit... but apparently you can only put a given file in one thread... sigh. If you're interested, I'm sure you can turn it up...

Okay, gotta go - machining a lens mount.

Cheers,

Ryan

Wayne Morellini · March 13th, 2007, 11:20 AM

Did I post this. Bugger, my head is still spinning after recent sickness.

Yes, I see your point, for the DOF adapter projects I did find some nice higher res Optical tapers you mentioned, with low light loss. For what I am looking for, the lens thing is probably better and cheaper. But thanks fro the idea. Fibre Optics also suffers some light scattering based on angle of entry and exit, so this solution would require mounting directly to the sensor, or the sensor faceplate, which could be a bit tricky.

There was somebody designing his own prism out of water, whatever happened to that?

Thanks

Wayne.

Ryan Damm · March 18th, 2007, 10:25 AM

Yeah, if you go with fiber optics it really must be glued to the sensor. A scary proposition, and it only works with single-chip cameras.

And lenses or not, you'll still need a scattering medium (cf previous optical-only discussion). That seems to be where most of the discussion is going on....

Oh, and I'm glad to hear you're feeling better - sounds like that illness really laid you low.

Rich Hibner · September 15th, 2007, 11:13 PM

reading through and seen someone posted this on another forum maybe it's worth checking out

http://www.abelcine.com/articles/ind...=170&Itemid=34

Jon Wolding · September 16th, 2007, 05:39 PM

I really want to be impressed by JVC's adapter, but I'm not. It'd be cheaper (and look about the same or better) to get an HPX500 w/CAC than to get the JVC kit w/ primes.

Rich Hibner · October 9th, 2007, 05:59 PM

this is my sort of optical solution

right click save as:
http://www.fileden.com/files/2007/1/...6/Optical2.mov

Mike Oveson · October 10th, 2007, 12:04 AM

Now Rich brings up an interesting point. The thread that this video comes from indicates that the person who created that clip used simply two 8x achromats and the 35mm lens. Having been off the adapter scene for some time I was rather shocked to read that, as my last impression when researching (and building) adapters was that this is impossible. How do you capture depth of field from an aerial image? If someone can break this down further from an optics standpoint I'd appreciate it.

Jon Wolding · October 10th, 2007, 09:59 PM

It's my understanding that an optical solution is possible for small CMOS chips... but not CCD chips.

Also, I'm curious about that video clip... I saw what looked like dust on the GG around 1:44, but maybe it was a compression artifact. What thread was this posted in? I want to know more about the setup.

Igor Babic · October 11th, 2007, 08:33 AM

http://www.dvxuser.com/V6/showthread.php?p=1067907

Rich Hibner · October 11th, 2007, 01:17 PM

Hey guys, Car3o=me. I made that clip.
The dust was on the lens. You can see it when her hand comes up and just a bit on her face. It might even be a scratch.

The setup was exactly as I described it. 2(8x achromats) and 1.5inch spacer then Nikon mount to lens. That's it. The lens I believe was a 28mm-78mm 3.5 zoom lens. The clip was actually a lot brighter than that. I'll see if I can find the original no CC to it.

At exactly 1:17 you can see the line which I think is a scratch or dust.

Igor Babic · October 11th, 2007, 02:24 PM

Can you post picture of your exact setup(spacers and step rings)? Are your achromats on opposite directions? Can you post link where you buy your 8x achromats? Is this same as your LI-35 setup but with only one Asian spacer? Are does achromats inside that tube like on your Li-35?
I have posted here your LI-35 picture from DVX forum, I hope I have'nt done something wrong.

PS
Why is so hard to focus with this setup?

Rich Hibner · October 11th, 2007, 06:14 PM

Here's a quick video of my adapter being put together and a video to follow it later with more tests. I hope this helps and it makes more want to find out how we can make this work better. I'm sure this is not a proper way to do it, but it seems to have sparked interest and that's a step closer to making a real one.

I get my achromats from this guy.

http://stores.ebay.com/COMA-Costruzi...QQftidZ2QQtZkm

right click save as:
http://www.fileden.com/files/2007/1/...6/QuickTut.mov
http://www.fileden.com/files/2007/1/...7;20Video1.mov

Igor Babic · October 12th, 2007, 02:19 AM

Thanks Rich, I already have all those stuff so I will try this at weekend.
It is very intresting.
One more question about your LS diffuser idea. Did you try to contact www.luminitco.com? They have 9 piece sample kit for 29$ (they send me an info about it today). If like you say this works like EE-a, this is also an exellent solution for gg type of adapter.

Matteo Pozzi · October 12th, 2007, 02:38 AM

hi Rich can you tell me what is the focal length and the diameter of the 8x achromat lens
many thanks
Matteo

February 22nd, 2007, 04:49 PM	#2
Ryan Damm Regular Crew Join Date: Dec 2005 Posts: 46	Telecentricity - the problem's in the aberrations It's actually pretty easy to make a lens system telecentric - all you have to do is put a single element lens in the path with the proper focal length. Essentially, the additional lens should be one focal length away from the exit pupil of the previous lens system. The ray bundles emerging from the second lens will now be parallel. They will also be severely aberrated. This is why optimized telecentric lenses have 50% to 300% more elements than a non-telecentric lens. In fact, 'doing it in stages' is exactly how most designs handle it: each element contributes some ray bending, and introduces/cancels some aberrations. This is not for the faint of heart; I have a friend who had to design a telecentric zoom lens in college without dedicated lens design software... he said it took a long, long time (sadistic professor, I suppose). Even using software it's daunting - that's why optical engineers are optical engineers. And could a still digital lens be used as a macro/relay? I don't see why not, but there may be issues I'm not aware of. Your point about the 'amount' of telecentricity is a good one - I believe the lenses are telecentric to avoid acceptance angle issues, which may be more forgiving than the beam splitting prisms on 3-CCD camcorders. If not, the only engineering challenge for you would be securing the proper flange focal length - unfortunately, it may be true that the flange focal length on the 4/3 lens is very short, shorter than the optical path distance from the CCD to the front of the prism block. You could extend that back focal length with a negative lens between the 4/3 lens and the prism block, but you'd change the FOV (essentially zooming in) and diverge the ray bundles - that is, you'd lose some telecentricity. This would also decrease the effective speed (f-stop) of the relay lens. And the previous point still holds - the relay lens would have to be awfully fast to get the DOF benefits of the front lens. Again, how fast depends on what your magnification ratio would be. (The magnification ratio is equal to the f-stop ratio; so if your chip is half the size of the target plane, your relay lens must be twice as fast - half the f-number - as the taking lens.) For my money, if you're going to be starting from scratch with a sensor, I'd look into fiber optic tapers. You'd still have to make the ray bundles somewhat telecentric - but probably a single-element or even achromat field lens would give you enough ray bending to get inside the acceptance cone of the taper. The advantages? No tricky aberrations or optical design software, and you would actually see a several stop gain (tapers are used in night vision for that very reason). I tried attaching a file (100r.pdf) which talks about tapers a little bit... but apparently you can only put a given file in one thread... sigh. If you're interested, I'm sure you can turn it up... Okay, gotta go - machining a lens mount. Cheers, Ryan

March 18th, 2007, 10:25 AM	#4
Ryan Damm Regular Crew Join Date: Dec 2005 Posts: 46	Tapers? Yeah, if you go with fiber optics it really must be glued to the sensor. A scary proposition, and it only works with single-chip cameras. And lenses or not, you'll still need a scattering medium (cf previous optical-only discussion). That seems to be where most of the discussion is going on.... Oh, and I'm glad to hear you're feeling better - sounds like that illness really laid you low.

October 10th, 2007, 12:04 AM	#8
Mike Oveson Major Player Join Date: Jan 2006 Location: Ogden, UT Posts: 349	Now Rich brings up an interesting point. The thread that this video comes from indicates that the person who created that clip used simply two 8x achromats and the 35mm lens. Having been off the adapter scene for some time I was rather shocked to read that, as my last impression when researching (and building) adapters was that this is impossible. How do you capture depth of field from an aerial image? If someone can break this down further from an optics standpoint I'd appreciate it. __________________ Endless Images

October 11th, 2007, 08:33 AM	#10
Igor Babic Regular Crew Join Date: Dec 2006 Location: Zagreb, Croatia Posts: 171	part is here http://www.dvxuser.com/V6/showthread.php?p=1067907

October 11th, 2007, 02:24 PM	#12
Igor Babic Regular Crew Join Date: Dec 2006 Location: Zagreb, Croatia Posts: 171	Can you post picture of your exact setup(spacers and step rings)? Are your achromats on opposite directions? Can you post link where you buy your 8x achromats? Is this same as your LI-35 setup but with only one Asian spacer? Are does achromats inside that tube like on your Li-35? I have posted here your LI-35 picture from DVX forum, I hope I have'nt done something wrong. PS Why is so hard to focus with this setup? Attached Images

March 13th, 2007, 11:20 AM	#3
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Did I post this. Bugger, my head is still spinning after recent sickness. Yes, I see your point, for the DOF adapter projects I did find some nice higher res Optical tapers you mentioned, with low light loss. For what I am looking for, the lens thing is probably better and cheaper. But thanks fro the idea. Fibre Optics also suffers some light scattering based on angle of entry and exit, so this solution would require mounting directly to the sensor, or the sensor faceplate, which could be a bit tricky. There was somebody designing his own prism out of water, whatever happened to that? Thanks Wayne.

September 15th, 2007, 11:13 PM	#5
Rich Hibner Major Player Join Date: Jan 2006 Posts: 218 Images: 1	reading through and seen someone posted this on another forum maybe it's worth checking out http://www.abelcine.com/articles/ind...=170&Itemid=34

September 16th, 2007, 05:39 PM	#6
Jon Wolding Regular Crew Join Date: Jan 2007 Location: Tampa, FL Posts: 195	I really want to be impressed by JVC's adapter, but I'm not. It'd be cheaper (and look about the same or better) to get an HPX500 w/CAC than to get the JVC kit w/ primes.

October 9th, 2007, 05:59 PM	#7
Rich Hibner Major Player Join Date: Jan 2006 Posts: 218 Images: 1	this is my sort of optical solution right click save as: http://www.fileden.com/files/2007/1/...6/Optical2.mov

October 10th, 2007, 09:59 PM	#9
Jon Wolding Regular Crew Join Date: Jan 2007 Location: Tampa, FL Posts: 195	It's my understanding that an optical solution is possible for small CMOS chips... but not CCD chips. Also, I'm curious about that video clip... I saw what looked like dust on the GG around 1:44, but maybe it was a compression artifact. What thread was this posted in? I want to know more about the setup.

October 11th, 2007, 01:17 PM	#11
Rich Hibner Major Player Join Date: Jan 2006 Posts: 218 Images: 1	Hey guys, Car3o=me. I made that clip. The dust was on the lens. You can see it when her hand comes up and just a bit on her face. It might even be a scratch. The setup was exactly as I described it. 2(8x achromats) and 1.5inch spacer then Nikon mount to lens. That's it. The lens I believe was a 28mm-78mm 3.5 zoom lens. The clip was actually a lot brighter than that. I'll see if I can find the original no CC to it. At exactly 1:17 you can see the line which I think is a scratch or dust.

October 11th, 2007, 06:14 PM	#13
Rich Hibner Major Player Join Date: Jan 2006 Posts: 218 Images: 1	Here's a quick video of my adapter being put together and a video to follow it later with more tests. I hope this helps and it makes more want to find out how we can make this work better. I'm sure this is not a proper way to do it, but it seems to have sparked interest and that's a step closer to making a real one. I get my achromats from this guy. http://stores.ebay.com/COMA-Costruzi...QQftidZ2QQtZkm right click save as: http://www.fileden.com/files/2007/1/...6/QuickTut.mov http://www.fileden.com/files/2007/1/...7;20Video1.mov Last edited by Rich Hibner; October 11th, 2007 at 07:11 PM.

October 12th, 2007, 02:19 AM	#14
Igor Babic Regular Crew Join Date: Dec 2006 Location: Zagreb, Croatia Posts: 171	Thanks Rich, I already have all those stuff so I will try this at weekend. It is very intresting. One more question about your LS diffuser idea. Did you try to contact www.luminitco.com? They have 9 piece sample kit for 29$ (they send me an info about it today). If like you say this works like EE-a, this is also an exellent solution for gg type of adapter.

October 12th, 2007, 02:38 AM	#15
Matteo Pozzi Regular Crew Join Date: Jul 2005 Location: Milan, Italy Posts: 169	hi Rich can you tell me what is the focal length and the diameter of the 8x achromat lens many thanks Matteo