Building DIY cylindrical lens? - Page 2 - AVS Forum | Home Theater Discussions And Reviews
Baselworld is only a few weeks away. Getting the latest news is easy, Click Here for info on how to join the Watchuseek.com newsletter list. Follow our team for updates featuring event coverage, new product unveilings, watch industry news & more!


Forum Jump: 
Reply
 
Thread Tools
post #31 of 60 Old 02-27-2012, 10:10 AM
Member
 
applesandrice's Avatar
 
Join Date: Apr 2004
Location: In my parents' basement.
Posts: 110
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
For more info about playing around with eyeglass lenses, check out the what Aussie Bob says in this post:

http://www.avsforum.com/avs-vb/showt...9#post16518239

He talks elsewhere about experimenting with these lenses, too, and he's a wealth of knowledge and information.
applesandrice is offline  
Sponsored Links
Advertisement
 
post #32 of 60 Old 02-27-2012, 12:01 PM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Aussie Bob is my new hero. Thank you for pointing this out. I've been searching for exactly this information for a long time.

It would seem (and I believe his description) that eyeglass lenses are never going to be the next Isco, but my next question is could they be better than a set of prisms. The need for an 83mm air gap might also pose problems with the size of these lenses and the throw ratios of typical projectors.
riskpeep is offline  
post #33 of 60 Old 02-27-2012, 12:19 PM
Scott Horton, techht.com
 
GetGray's Avatar
 
Join Date: Aug 2004
Location: Mid-South USA
Posts: 5,569
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 93 Post(s)
Liked: 83
He hasn't participated here in a while but you can contact him via www.xeitoptics.com if you want to chat about the technical details on lens design.
GetGray is offline  
post #34 of 60 Old 02-27-2012, 12:43 PM
Advanced Member
 
mrfattbill's Avatar
 
Join Date: Aug 2009
Posts: 902
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 46 Post(s)
Liked: 70
Quote:
Originally Posted by riskpeep View Post

Aussie Bob is my new hero. Thank you for pointing this out. I've been searching for exactly this information for a long time.

It would seem (and I believe his description) that eyeglass lenses are never going to be the next Isco, but my next question is could they be better than a set of prisms. The need for an 83mm air gap might also pose problems with the size of these lenses and the throw ratios of typical projectors.

I believe he said that you can't just order the correct size needed out of a catalog but it would have to be a custom cut piece which will add greatly to the price, at least when he made the post (09).

Bill
mrfattbill is offline  
post #35 of 60 Old 02-27-2012, 01:27 PM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Yeah, I've noticed that about the largest lens blanks you can find are in the 100mm range, most are 70-80mm range.

One of the reasons I started looking at eyeglass lenses, is that outside of those, for high quality optics, most parts are in the 25mm or smaller sizes. The prices get ridiculous if you get up much past that.

I still haven't given up on the idea of using lens blanks, but reading aussie bob's posts have given me pause. I had heard that lens blanks aren't ground to the same precision as your average camera or projector lens, but to have it confirmed pushes me back a step.

My current like of research (if you can call google searching research) is to see if I can quantify the parameter we care about (surface quality?, surface deviation?), and then compare various kinds of lens materials to that to see how close I can find. Many firms now tout 'High-Definition' lenses that I think might be market speak for better surface quality or lower surface deviation. I'm still trying to figure out the market speak from the engineer speak (which is what I want).

I have no allusions that I'm going to create a $50 ISCO lens. But I'd like to think that we should be able to outdo your typical prism lens.
riskpeep is offline  
post #36 of 60 Old 02-27-2012, 04:04 PM
AVS Special Member
 
CAVX's Avatar
 
Join Date: Aug 2005
Location: Brisbane, Australia
Posts: 8,417
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 39 Post(s)
Liked: 55
Quote:
Originally Posted by riskpeep View Post

Mark,

Thank you for your reply, this is helpful. Is the distance between the lenses particularly relevant, or just the alignment between them?

Spacing is absolutely critical and why I now have a massive threaded focus ring on the MK5.



Lets see if this gets me another suspension . At the distance it is set in this photo (recessed), the image is not in focus at my TR and projection distance. To be in focus, the threaded centre needs to protrude about 12mm extending the depth of the lens to about 140mm.

Quote:
I've been doing some additional research on this and what I know so far is that eyeglass blanks are measured in diopters. At the risk of being pedantic, a diopter (http://en.wikipedia.org/wiki/Dioptre) is a measure of the optical power of a lens and is the reciprocal of the focal length, so a 3 diopter lens would have a focal length of 1/3 of a meter. Lenses can be found in 0.25 increments of diopters.

Eye glass blanks are typically CR39 plastic these days and you should be able to specify a cylindrical cut at a specific diopter. As you see from the first image I attached, you need both a concave and convex lens pair. They can be "round" or "square". I've drawn then "square" because I could not get Sketch Up to allow me to cut the profile round after cutting the basic shape. Something about not being able to shift curved surfaces?

Quote:
The interesting bit is that for an eyeglass prescription (and thus for readily available lenses), you have a spherical measure (which is a general magnification), and a cylindrical measure. We only care about cylindrical power, so the appropriate lenses with the correct cylinder rating is what we'd need.

A blank should be a small cylinder - R=40mm H=10~15mm. They cut them to be spherical (both x and y axis) where a cylindrical lens is just one axis.

Quote:
My understanding is that for an Anamorphic lens, we want a 1.33x expansion lens which should be achievable with two lenses having the right ratio between their Diopter ratings. I found one person on the web who'd had success with experimenting at an anamorphic stretch using a -6 cylinder lens paired with a +4.5 cylinder lens, but he was using opthalmic test lenses with a diameter of ~1" and thus couldn't pass the whole beam of his projector.

I read that too and where 6/4.5=1.33. There are many other combinations that can give 1.33x

You should be able to use any two lenses with the correct ratio between them, so if I have this right, you could use a -6 & a +4.5 cylinder lens, or a -3 and a +2.25 lens.

Quote:
If we can find someone with access to eyeglass lens blanks and a projector, it would seem a simple test to put two lenses in front of a projector to see if we're approaching this problem right. Obviously you'd need to get the orientation right (being cylindrical lenses, they do have a directionality, and between the two lenses the directionality would need to be exactly lined up.

You might have to go to an optometrist and place an order for these lenses. If they can cut the precise multi-aspheric curves of corrective eye wear, I am sure that they can cut the simple single radii you need - especially because it is in just one direction.

Quote:
Yeah, I've noticed that about the largest lens blanks you can find are in the 100mm range, most are 70-80mm range.

One of the reasons I started looking at eyeglass lenses, is that outside of those, for high quality optics, most parts are in the 25mm or smaller sizes. The prices get ridiculous if you get up much past that.

Welcome to my world. I still have the very first plano convex lens I bought back in 2006. 25mm, BK7 singlet and cost me just $15USD. I do believe it is the 6D as described in that thread at DIY Audio, but I could not find a 4.5D concave to match for an anamorphic pair.

Quote:
I still haven't given up on the idea of using lens blanks, but reading aussie bob's posts have given me pause. I had heard that lens blanks aren't ground to the same precision as your average camera or projector lens, but to have it confirmed pushes me back a step.

Whist his posts are very informative, he tends to do that. I wonder how different thing would have turned out if I actually took his "don't bother" advise back in 2005?

Quote:
My current like of research (if you can call google searching research) is to see if I can quantify the parameter we care about (surface quality?, surface deviation?), and then compare various kinds of lens materials to that to see how close I can find. Many firms now tout 'High-Definition' lenses that I think might be market speak for better surface quality or lower surface deviation. I'm still trying to figure out the market speak from the engineer speak (which is what I want).

I have no allusions that I'm going to create a $50 ISCO lens. But I'd like to think that we should be able to outdo your typical prism lens.

Commercial optical surface finish is scratch and dig 60/40. That means the largest scratch on the surface of the glass should be no greater than 60 x 40 microns. There is not even a published spec for those trophies. All I ever got from the manufacture was something about a 4 step polishing process.

There is nothing wrong with SD 60/40 and in fact it is the spec used by at least 2 anamorphic lenses available today. Sure you ask for a better surface finish, but your lens set just cost you triple.
LL

Mark Techer

I love my Constant Image Height system!
CAVX is offline  
post #37 of 60 Old 02-29-2012, 02:05 PM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Quote:


Eye glass blanks are typically CR39 plastic these days and you should be able to specify a cylindrical cut at a specific diopter. As you see from the first image I attached, you need both a concave and convex lens pair.

Yeah CR39 is very common. Here in the US, you can get also find glass as well as a number of high index plastics although these cost more and can be special order items. If we can work out the finish quality and light path issues, my thought would be to use different materials to see what we could do with correcting CA. (that is a problem for another day)

Back to the lens design. If I'm doing my math right, it looks like the following lens combinations would work. I stopped at -10 because most lens blanks aren't available past -10, and for most retailers anything beyond -3 is a special order item.

Lens 1 Lens 2 FL1(mm) FL2(mm) Distance(mm)
-1 0.75 -1000 1333.333333 333.3333333
-2 1.5 -500 666.6666667 166.6666667
-3 2.25 -333.3333333 444.4444444 111.1111111
-4 3 -250 333.3333333 83.33333333
-6 4.5 -166.6666667 222.2222222 55.55555556
-8 6 -125 166.6666667 41.66666667
-10 7.5 -100 133.3333333 33.33333333

From Aussie Bob's post, I know that we need to space the lenses based on their focal lengths (FL1 & FL2 in the table), with the distance being the Distance column.

Since we have such small lenses to work with, I think we'd want to minimize the required distance between them to make sure that the light stays within the usable area of the lenses. A longer lens is going to need bigger glass to capture the 'fast light' coming out of a projector.

TR will also come into play here. I think I've got some math ahead of me to figure out what the ratio of input aperture to exit apertures are indicated by various throw ratios. It is possible, perhaps even likely that even for a long TR, that 80mm on the exit aperture won't do it even for a very short lens.

I've been told that most blanks are 80mm, so I'm going to assume that in my entry/exit aperture analysis.

SOHCAHTOA here I come.

Quote:


Commercial optical surface finish is scratch and dig 60/40. That means the largest scratch on the surface of the glass should be no greater than 60 x 40 microns.

It sounds like this would be a surface quality measure. Is there a surface deviation metric as well that is typical? (i.e. a measure of how true the shape of the lens is)
riskpeep is offline  
post #38 of 60 Old 02-29-2012, 03:25 PM
AVS Special Member
 
CAVX's Avatar
 
Join Date: Aug 2005
Location: Brisbane, Australia
Posts: 8,417
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 39 Post(s)
Liked: 55
Quote:
Originally Posted by riskpeep View Post

Yeah CR39 is very common. Here in the US, you can get also find glass as well as a number of high index plastics although these cost more and can be special order items. If we can work out the finish quality and light path issues, my thought would be to use different materials to see what we could do with correcting CA. (that is a problem for another day)

AFAIK, the only way to correct for CA is to use a achromatic doublet which is two lenses made of different glass types that are cemented together to form a single lens.

Quote:


Back to the lens design. If I'm doing my math right, it looks like the following lens combinations would work. I stopped at -10 because most lens blanks aren't available past -10, and for most retailers anything beyond -3 is a special order item.

Lens 1 Lens 2 FL1(mm) FL2(mm) Distance(mm)
-1 0.75 -1000 1333.333333 333.3333333
-2 1.5 -500 666.6666667 166.6666667
-3 2.25 -333.3333333 444.4444444 111.1111111
-4 3 -250 333.3333333 83.33333333
-6 4.5 -166.6666667 222.2222222 55.55555556
-8 6 -125 166.6666667 41.66666667
-10 7.5 -100 133.3333333 33.33333333
From Aussie Bob's post, I know that we need to space the lenses based on their focal lengths (FL1 & FL2 in the table), with the distance being the Distance column.

In theory, all of those should give you 1.33x or 0.75x expansion based on how the light enters/exits the A-Lens.

Quote:


Since we have such small lenses to work with, I think we'd want to minimize the required distance between them to make sure that the light stays within the usable area of the lenses. A longer lens is going to need bigger glass to capture the 'fast light' coming out of a projector.

Bear in mind, I have a working example of an anamorphic lens made from what I believe are these eye glass blanks. The front lens is just 80mm diameter. The rear is closer to 60mm. The lens works at longer throws like 2.1.

Mark Techer

I love my Constant Image Height system!
CAVX is offline  
post #39 of 60 Old 02-29-2012, 07:28 PM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Quote:


Bear in mind, I have a working example of an anamorphic lens made from what I believe are these eye glass blanks. The front lens is just 80mm diameter. The rear is closer to 60mm. The lens works at longer throws like 2.1.

ORLY! How well does it work? From Aussie Bob's comments I'd think not so well, but I'm thinking he might be a bit of a purist. I've no doubt that your custom lenses are better, but how about as compared to your typical prism lens?
riskpeep is offline  
post #40 of 60 Old 02-29-2012, 08:34 PM
AVS Special Member
 
CAVX's Avatar
 
Join Date: Aug 2005
Location: Brisbane, Australia
Posts: 8,417
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 39 Post(s)
Liked: 55
Quote:
Originally Posted by riskpeep View Post

ORLY! How well does it work? From Aussie Bob's comments I'd think not so well, but I'm thinking he might be a bit of a purist. I've no doubt that your custom lenses are better, but how about as compared to your typical prism lens?

AB has said allot of things

Mark Techer

I love my Constant Image Height system!
CAVX is offline  
post #41 of 60 Old 03-25-2012, 08:51 AM
AVS Special Member
 
AVWERKS's Avatar
 
Join Date: Jan 2003
Location: Southern cal
Posts: 1,270
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 10
Is there anything from an old CRT lense that could be used? They are certainly big enough and looking at a cross section of a NEC XG lens from a brochure for example, I see every kind of shape imaginable in one lens. Some might be plastic though, but wouldn,t know until inspection. I have a few different lens available from different CRT projectors available also. Certainly easy enough to take apart.

So What are the physical specs of the lens you have made?
This might be a moot point if I understand the real specs as being either a concave or convex in one axis only of each lens?
Maybe one of these crt lenses has what it takes or maybe a certain mix to get there?
Regards
David

Talk is cheap because supply exceeds demand
AVWERKS is offline  
post #42 of 60 Old 03-25-2012, 02:44 PM
AVS Special Member
 
CAVX's Avatar
 
Join Date: Aug 2005
Location: Brisbane, Australia
Posts: 8,417
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 39 Post(s)
Liked: 55
Not really as the lenses in those will all be spherical and you need cylindrical for this.

Mark Techer

I love my Constant Image Height system!
CAVX is offline  
post #43 of 60 Old 10-14-2012, 06:47 AM
Senior Member
 
Gradius2's Avatar
 
Join Date: Mar 2006
Location: Iquique, Chile
Posts: 412
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 19 Post(s)
Liked: 14
I wonder how much those lens from ISCO costs directly from manufacturer.

I can get them directly in Germany, this is why.

[]s,
Fernando
Gradius2 is offline  
post #44 of 60 Old 10-14-2012, 08:58 AM
Scott Horton, techht.com
 
GetGray's Avatar
 
Join Date: Aug 2004
Location: Mid-South USA
Posts: 5,569
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 93 Post(s)
Liked: 83
Quote:
Originally Posted by Gradius2 View Post

I wonder how much those lens from ISCO costs directly from manufacturer.

I can get them directly in Germany, this is why.
Fernando: Isco brand is only sold to OEM customers, myself included. Schneider branded lenses are sold to the dealer channel. None are sold to end users directly. That said, if you Email me I can price the Isco for you. We arrange for shipping to or inside Germany (and worldwide) regularly. scott at techht dot com.
GetGray is offline  
post #45 of 60 Old 02-12-2015, 11:45 PM
Newbie
 
Join Date: Feb 2015
Posts: 8
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 0
Hey guys!

Not sure if anyone is still here but could anyone explain to me what the values in the Lens 1 and Lens 2 represent?
I would really appreciate it!
PIanet is offline  
post #46 of 60 Old 03-05-2015, 11:16 PM
AVS Special Member
 
CAVX's Avatar
 
Join Date: Aug 2005
Location: Brisbane, Australia
Posts: 8,417
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 39 Post(s)
Liked: 55
Quote:
Originally Posted by riskpeep View Post
Quote:

Bear in mind, I have a working example of an anamorphic lens made from what I believe are these eye glass blanks. The front lens is just 80mm diameter. The rear is closer to 60mm. The lens works at longer throws like 2.1.


ORLY! How well does it work? From Aussie Bob's comments I'd think not so well, but I'm thinking he might be a bit of a purist. I've no doubt that your custom lenses are better, but how about as compared to your typical prism lens?
The little cylindrical lens kills the performance of a prism lens, but is no where as good as my MK5. That is for certain. I think at some point this little lens has been dropped and the vertical alignment I spoke several times about in this thread is now out. It won't focus 100% in my system at 1080 rez. I think it would be fine for 480 and maybe 720 though.

Mark Techer

I love my Constant Image Height system!
CAVX is offline  
post #47 of 60 Old 03-05-2015, 11:43 PM
Member
 
Aussie Bob II's Avatar
 
Join Date: Oct 2012
Location: Sydney, Australia
Posts: 76
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 17
I'm blushing with embarrassment. All this talk about a few words I wrote years ago AND I MISSED IT!!!

Quote:
I'm an optometrist so I can easily get high grade glass cylinder lenses, with anti-reflection, for very little cost.
Anti reflection coatings are about 10% of the problem. Eyeglass lens blanks are not color corrected. Hence, if you build an anamorphic couplet using CR39, be prepared for all the colors of the rainbow from about halfway out to the edge of the screen.

Eyeglass lenses are meniscus lenses. Not suitable for anamorphic couplets. You will observe, not only color aberration, but also blurring anywhere but near the center of the image, getting worse out towards the edge of the screen.

Quote:
No one on here will give you any info on how to build one, or more importantly, the optical settings due to the fact that they spent a bunch of time figuring it out and are now selling the products and don't want to lose out to DIY folk making their own. I personally think the cylinder lenses are marked up A LOT, but then again that's the market price until someone comes out with a cheaper version. But i could be wrong.
You turned out to be wrong about that. I gave out the information because there's a lot more to building an anamorphic lens than just the ratios of the focal lengths.


Quote:
If building your own lens was this costly ("$$$$")and difficult to build your own cylinder lens then they would easily give that info out because they know you would fail and/or give up due to start up costs. So I'm thinking its not as bad as it sounds.
It's much, MUCH worse than it sounds. And incredibly expensive to prototype large chunks of anamorphic glass.

Quote:
1/100ths of a degree would only be a measurement important to NASA definitely not to the system.
Wrong. 1/100th of a degree is a minimum for 1080p, and not good enough for 4K. Tiny miscalibrations can be fixed by counter-rotating the other lenses in the set in the opposite direction. But when it boils down to it the accuracy still has to be 1/100th of a degree, even if it's a kluge to fix something out of alignment. When you counter-rotate lenses to fix a problem you also skew the image. It's no longer rectangular, but rhomboid.

Quote:
Nobody can correct it that close, in fact that amount of detail is insignificant for a optical lens for a HT.
I do, routinely. In fact our correction is better than 1/150th of a degree. It's easy (in a relative sense) when you invest in calibration equipment.

Quote:
IF that were true then the guy that had the -6.00 amount of astigmatism in his glasses would not be able to see anything with glasses as all glasses are corrected on intervals of 5 degrees, no where near 1/100th of a degree.
You can't really measure whether something appears to be "level" to a subjective human perception, certainly not down to 1/100th of a degree. You CAN measure - very accurately - whether an anamorphically expanded image is level. The human eye, and its perception, are a completely different ball game to measurable, verifiable standards applied to scientific instruments.

Quote:
If you make a system corrected to that detail and that is important to you then you need to get your eye prescription updated, for changes, every month!
Fatuous comment. The human eye is much more forgiving than a laser-calibrated microscope used for precise collimation.

Quote:
It generally takes 4 cylinder lenses (cylinder lenses are the type used to correct peoples vision when they have astigmatism: ie eye shaped like an american football) The more lenses used, the more optical clarity that can be corrected for. But I haven't actually worked it out yet but the materials aren't that expensive nor do you need to work for NASA.
Eyeglass lenses are plastic, made to dramatically lower standards than very expensive cylinder lenses. The human eye can adapt focus and to some extent anastigmatic error. Projectors and projection lenses cannot. If they're wrong, they're wrong.

I now have hundreds of lens customers, optical freaks, who do "hi-fi fanatic" in their spare time. These people are unforgiving of the slightest flaw in their precious picture. they have invested thousands, tens of thousands and sometime hundreds of thousands of dollars in their theaters and production studios NOT to have it degraded by a dud lens (which is, after all, the final optical summation of their entire investment, as it is the last thing the image passes through before it hits the screen). There are no shortcuts to excellence. They need the best contrast, the best sharpness, the lowest distortion and the brightest images, virtually indistinguishable from the original, as possible. Some are private customers, some are in government, some in defence and film production. But they all share a common trait: they don't invest in garbage, and they don't like bulls*it.

Someone above wrote that I say a lot. Not so much lately, but if I do say anything it is this: there is no shortcut in optics. There is no dirty little secret that optical engineers keep to themselves, to chuckle about when they're having a beer with the other optical engineers in on the scam. There is only hard work, spiced with a lot of luck and many hundreds of hours of toil. You can't use eyeglass blanks, you can't bend a plastic mirror (another favourite) and you can't buy cheap trophy prisms to achieve the near-to-perfect results that customers demand. You have to work at it, keeping your environment scrupulously clean, ordered and as pristine as the day you first put it together. If you don't believe me, try the cheap approach. It won't work, but you MAY learn (the hard way) why it doesn't.
Aussie Bob II is offline  
post #48 of 60 Old 03-06-2015, 05:25 AM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Quote:
Originally Posted by PIanet View Post
Hey guys!

Not sure if anyone is still here but could anyone explain to me what the values in the Lens 1 and Lens 2 represent?
I would really appreciate it!
Those measures are in Dioptres. They are a measure used in the eyeglass industry to characterize the power of a lens. In this case, you'd want a 0 sphere, X cylinder, 0 axis lens, where the X is the number from the table. You don't want any sphere because we're only trying to create expansion in one direction. Normally w/ cylinder measures, you have to specify an axis too, but that is if they are going to be put in eyeglasses, so you can just use 0. The axes of the two lenses would need to line up, however.

http://en.wikipedia.org/wiki/Dioptre

Rob

Last edited by riskpeep; 03-06-2015 at 05:29 AM.
riskpeep is offline  
post #49 of 60 Old 03-06-2015, 05:38 AM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Quote:
Originally Posted by Aussie Bob II View Post
I'm blushing with embarrassment. All this talk about a few words I wrote years ago AND I MISSED IT!!!



Anti reflection coatings are about 10% of the problem. Eyeglass lens blanks are not color corrected. Hence, if you build an anamorphic couplet using CR39, be prepared for all the colors of the rainbow from about halfway out to the edge of the screen.

Eyeglass lenses are meniscus lenses. Not suitable for anamorphic couplets. You will observe, not only color aberration, but also blurring anywhere but near the center of the image, getting worse out towards the edge of the screen.



You turned out to be wrong about that. I gave out the information because there's a lot more to building an anamorphic lens than just the ratios of the focal lengths.




It's much, MUCH worse than it sounds. And incredibly expensive to prototype large chunks of anamorphic glass.



Wrong. 1/100th of a degree is a minimum for 1080p, and not good enough for 4K. Tiny miscalibrations can be fixed by counter-rotating the other lenses in the set in the opposite direction. But when it boils down to it the accuracy still has to be 1/100th of a degree, even if it's a kluge to fix something out of alignment. When you counter-rotate lenses to fix a problem you also skew the image. It's no longer rectangular, but rhomboid.



I do, routinely. In fact our correction is better than 1/150th of a degree. It's easy (in a relative sense) when you invest in calibration equipment.



You can't really measure whether something appears to be "level" to a subjective human perception, certainly not down to 1/100th of a degree. You CAN measure - very accurately - whether an anamorphically expanded image is level. The human eye, and its perception, are a completely different ball game to measurable, verifiable standards applied to scientific instruments.



Fatuous comment. The human eye is much more forgiving than a laser-calibrated microscope used for precise collimation.



Eyeglass lenses are plastic, made to dramatically lower standards than very expensive cylinder lenses. The human eye can adapt focus and to some extent anastigmatic error. Projectors and projection lenses cannot. If they're wrong, they're wrong.

I now have hundreds of lens customers, optical freaks, who do "hi-fi fanatic" in their spare time. These people are unforgiving of the slightest flaw in their precious picture. they have invested thousands, tens of thousands and sometime hundreds of thousands of dollars in their theaters and production studios NOT to have it degraded by a dud lens (which is, after all, the final optical summation of their entire investment, as it is the last thing the image passes through before it hits the screen). There are no shortcuts to excellence. They need the best contrast, the best sharpness, the lowest distortion and the brightest images, virtually indistinguishable from the original, as possible. Some are private customers, some are in government, some in defence and film production. But they all share a common trait: they don't invest in garbage, and they don't like bulls*it.

Someone above wrote that I say a lot. Not so much lately, but if I do say anything it is this: there is no shortcut in optics. There is no dirty little secret that optical engineers keep to themselves, to chuckle about when they're having a beer with the other optical engineers in on the scam. There is only hard work, spiced with a lot of luck and many hundreds of hours of toil. You can't use eyeglass blanks, you can't bend a plastic mirror (another favourite) and you can't buy cheap trophy prisms to achieve the near-to-perfect results that customers demand. You have to work at it, keeping your environment scrupulously clean, ordered and as pristine as the day you first put it together. If you don't believe me, try the cheap approach. It won't work, but you MAY learn (the hard way) why it doesn't.
Aussie Bob,

While I understand that most eyeglass lenses are made from CR39, it is possible here in the US (can't speak about elsewhere) to get actual glass lenses as well. Depending on where you buy from, they advertise it as Crown Glass, the same material used in many optics lenses, yes?

As for CA, yes, this is something that would need to be considered, but my thought was to see if we could use different materials for the lens pairs to correct some of the issues. I understand that correcting CA is a complicated problem that might not be possible with just two lenses and limited ability to control FL.

My goal here isn't to shortcut and beat Issco (even trying would be a particular form of hubris IMO), but rather to be better than prisms, which are the state of the art in the DIY community right now. I don't think people who've spent thousands on the rest of the system are the buyers here, it is more likely the guy who bought a thousand dollar projector off Amazon and wants to go one step farther in his basement.

R
riskpeep is offline  
post #50 of 60 Old 03-06-2015, 06:19 AM
Scott Horton, techht.com
 
GetGray's Avatar
 
Join Date: Aug 2004
Location: Mid-South USA
Posts: 5,569
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 93 Post(s)
Liked: 83
Quote:
Originally Posted by riskpeep View Post
My goal here isn't to shortcut and beat Issco (even trying would be a particular form of hubris IMO), but rather to be better than prisms, which are the state of the art in the DIY community right now. I don't think people who've spent thousands on the rest of the system are the buyers here, it is more likely the guy who bought a thousand dollar projector off Amazon and wants to go one step farther in his basement.
That guys needs to zoom. Or stick with 16:9. If you want acceptable scope presentation, it comes with an admission price. And as the new XEIT 4k improves on Isco and relatively speaking is a huge deal in the US through me, that ticket isn't that expensive IMO. If you are using a $1k PJ expecting an acceptable scope image from a home-made lens, your expectations are unrealistic.
GetGray is offline  
post #51 of 60 Old 03-06-2015, 07:24 AM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Quote:
Originally Posted by GetGray View Post
That guys needs to zoom. Or stick with 16:9. If you want acceptable scope presentation, it comes with an admission price. And as the new XEIT 4k improves on Isco and relatively speaking is a huge deal in the US through me, that ticket isn't that expensive IMO. If you are using a $1k PJ expecting an acceptable scope image from a home-made lens, your expectations are unrealistic.
<sigh>

This is the DIY forum isn't it?
riskpeep is offline  
post #52 of 60 Old 03-06-2015, 08:12 AM
Scott Horton, techht.com
 
GetGray's Avatar
 
Join Date: Aug 2004
Location: Mid-South USA
Posts: 5,569
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 93 Post(s)
Liked: 83
Yes, but I'd be discouraging if someone was asking how to DIY their own 3-chip DLP projector too.

Nothing would make me happier than to have an inexpensive A-lens available to me or the masses, I mean that sincerely. The ones I sell have very little (sometimes no) markup. They are a necessary evil to sell the CineSlide that I make. But I've seen some pretty ghetto solutions with prisims and I would not recommend those to anyone. If you can work some magic and produce a cheap cylindrical lens that works, I'll be the first one to order a few cases of them. But, because I know all the players in the A-lens world, and know what the costs are, why they are, and what the limitations are, I'm confident it would be magic. Don't mean to be discouraging for the sake of being negative in the DIY forum. All that said, even if you make one work that works poorly (or not), I'm still interested in watching. (I like technical magic shows)

Last edited by GetGray; 03-06-2015 at 08:25 AM.
GetGray is offline  
post #53 of 60 Old 03-06-2015, 08:23 AM
Member
 
riskpeep's Avatar
 
Join Date: Sep 2007
Posts: 17
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 10
Quote:
Originally Posted by GetGray View Post
Yes, but I'd be discouraging if someone was asking how to DIY their own 3-chip DLP projector too.

Nothing would make me happier than to have an inexpensive A-lens available to me or the masses, I mean that sincerely. The ones I sell have very little (sometimes no) markup. They are a necessary evil to sell the CineSlide that I make. But I've seen some pretty ghetto solutions with prisims and I would not recommend those to anyone. If you can work some magic and produce a cheap cylindrical lens that works, I'll be the first one to order a few cases of them. But, because I know all the players in the A-lens world, and know what the costs are, why they are, and what the limitations are, I'm confident it would be magic. Don't mean to be discouraging for the sake of being negative in the DIY forum. All that said, even if you make one work that works poorly (or not), I'm still interested in watching.
Well that is good to hear.

And that is exactly the thread I'm trying to pull. I've no doubt that a purpose built lens will look and function better than anything I'm suggesting. I'm trying to answer the question of if a set of these lenses might be better than the 'ghetto' prism lenses. Even if it isn't good, is it better than the currently bad thing that people are doing?

R
riskpeep is offline  
post #54 of 60 Old 03-06-2015, 02:39 PM
Member
 
Aussie Bob II's Avatar
 
Join Date: Oct 2012
Location: Sydney, Australia
Posts: 76
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 17
Quote:
And that is exactly the thread I'm trying to pull. I've no doubt that a purpose built lens will look and function better than anything I'm suggesting. I'm trying to answer the question of if a set of these lenses might be better than the 'ghetto' prism lenses. Even if it isn't good, is it better than the currently bad thing that people are doing?
Eyeglass lenses, even made from glass still won't cut it.

The reasons are several:

1. Color Aberration correction
2 lenses in an enclosure aren't enough to correct fully for color aberration. You need to match the color correction properties with the focal lengths, focal point and refraction characteristics of the glass you're using. Using two types of glass seems like a possible path to a solution, but it only seems that way

The kinds of glass or plastic available for eyeglass lenses are restricted to just a few indices of refraction. Usually higher indices are used to make eyeglass lenses thinner because they bend the light more than lower indices, don't need as much curvature, and thus can be thinner (especially at the edges). These are designed to work well with the human eye, but a human eye is not a projector lens. The odds of finding a matching pair of indices of refraction in conjunction with the diopter powers and air gaps required are vanishingly slim.

2. The shape of eyeglass lenses is all wrong
Eyeglass lenses typically are meniscus in form. For example, negative lenses (diverging lenses) have a spherical outside curvature and a cylindrical inside curvature (inside = "closest to the eye") in the same direction. Sure you can purchase the right diopter strengths out of a catalog, but the diopter rating is only a guide. For any specific focal length you can have lenses in an infinite number of shapes. Using a lens formula to calculate focal length from the two curvatures, the material's index of refraction and the thicknesses gives a definite focal length result, but if the shape of the lenses is wrong for the application (i.e. projection, not human eye correction) you may achieve some resolution in the middle but the image from there out to the sides of the screen will suffer greatly.

Having said that, we employ one large meniscus-shaped lens in our product, but it's shape is carefully calculated and controlled. It is matched against other more conventionally shaped lenses in the glass array to reduce distortion. An all-meniscus device will not produce particularly watchable images.

3. Quality of eyeglass lenses
The quality of eyeglass lenses is iffy. They have quite slack tolerances compared to custom ground lenses, and the materials can contain straie, solidified ebbs and flows in the actual lens material which locally alter focal length in just one spot. This comes from relatively less than perfectl over the material when it is cooling from molten to solid states. These are incredibly difficult to detect with an eyeball inspection, but when you're enlarging an image to the size of a wall they exhibit prominently. I even get them in the lenses we have made for us professionally, and they have to be discarded.


4. Size of eyeglass lenses
Most eyeglass lenses are 80mm in diameter, or close to it. This is too small for everything but the higher-end throw ratios.

5. Comparison against prismatic designs
A prismatic design will trump an eyeglass lens design, given good quality in both constructions. Prisms are far less complicated than eyeglass lenses, having only flat surfaces, and in this case that's a plus. However, even a good prism design is awful compared to a properly designed cylindrical lens kit.

6. The Dirty Little Secret that optical engineers are keeping from their mug customers
There isn't one. There is no cheap way of getting anything near 1080p quality from a projection system, much less 4K quality, if you want to use either prisms or (especially) eyeglass lenses. You'd be lucky to get even acceptable You Tube quality. All kinds of claims are made that this is possible. All of them are false. There are no shortcuts. If you want to resolve pixels down to half-a-millimeter size on a 3000 millimeter wide screen - that means resolving pixel tiling at around 0.1 millimeter size clearly - you have to use a properly designed and calibrated anamorphic lens.

7. Do it for fun?
Sure, but please don't then claim you've reinvented the science of optics and can produce an anamorphic lens for $100 that even 10% matches a professional, purpose-designed product. What fools a lot of people in this area is not having a comparison. At first your DIY lens will look fantastic, from a distance, but as you move closer to the screen you'll start to see aberrations, out of focus areas, color casts and a lot more bad things. Then put one of our Xeit lenses in front of your projector - with every pixel sharp, no color aberration, deep contrast, low distortion and no compromise - and you'll see the difference. You're either into high definition Home Cinema or you're just having some fun.
Aussie Bob II is offline  
post #55 of 60 Old 03-08-2015, 06:22 PM
Advanced Member
 
Metric's Avatar
 
Join Date: Jun 2001
Posts: 852
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 28 Post(s)
Liked: 17
Quote:
Originally Posted by GetGray View Post
That guys needs to zoom. Or stick with 16:9. If you want acceptable scope presentation, it comes with an admission price. And as the new XEIT 4k improves on Isco and relatively speaking is a huge deal in the US through me, that ticket isn't that expensive IMO. If you are using a $1k PJ expecting an acceptable scope image from a home-made lens, your expectations are unrealistic.

Yeah that's a pretty rude elitist statement... plenty of us are trying to push the envelope with what we have.

I want to die in my sleep like grandpa... not kicking and screaming like the people in his car.
Metric is offline  
post #56 of 60 Old 03-08-2015, 07:20 PM
Member
 
Aussie Bob II's Avatar
 
Join Date: Oct 2012
Location: Sydney, Australia
Posts: 76
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 17
Quote:
...plenty of us are trying to push the envelope with what we have.
It seems you won't be discouraged, so go ahead. Get Gray's statement was not elitist. It was a statement of fact.

Go ahead and try some eyeglass elements. You may even think you've discovered the Holy Grail. But unfortunately there is no el cheapo way of doing that. There are no dirty little secrets optical engineers keep to themselves.

You need to consider why projector manufacturers use multi-element lenses in their machines. Ditto for camera makers. It's not just to get the zooming right, but also to correct for color aberration, distortion, astigmatism, uneven focus, and literally thousands of other things that Thin Lens (or Ideal Lens) theory doesn't deal with. For every focal length there is literally an infinite number of curvatures, shapes, styles and materials from which to construct a lens of that particular focal length. Combining this with yet more lenses complicates the equations necessary to be solved by a further infinite amount. Yet there is a relatively small (microscopically so) sub-set of the ostensibly possible solutions that will come anywhere near working. Of these, only a further small sub-set will work at any throw distance or with any projector under any circumstances.

What you want to do is to put a cheap plastic (or glass, it doesn't matter) couplet of lenses in front of these pieces of complex optics and gain an acceptable image from them. It won't, and cannot happen, but never let it be said you died wondering.

Thin Lens theory will tell you that to attain a 1.33x expansion you need to use cylindrically-curved lenses - one diverging and one converging - whose focal lengths are in the ratio of 4:3 (1.33:1). The air gap between these two lenses will be the difference between their focal lengths in whatever units you are employing. So, a 400mm focal length converging lens, coupled with a -300mm focal length diverging lens will have a nominal spacing of 100mm (this will certainly need to be adjusted, so make allowances for that). The table above (post #37 ) seems pretty accurate on the range of theoretical values possible with standard optical blanks. I used a -4 diopter and a +3 diopter couplet (-250mm and 333mm focal lengths with 83.3mm spacing) because that was nice and short, but looked like it might accommodate the beam width (with the lenses both being 80mm in diameter). Any shorter focal lengths in the proper ratio will introduce too much distortion, although your rig will be shorter. Any longer will be too long, and won't fit the beam due to the length of the tube you'll have to use. The diverging lens (positive focal length) goes closest to the projector.Both lenses have their unpowered side (i.e the face that would be away from the eye) on the outside.

There you go, there's your primer on how to use optical lenses for an anamohpic attachment.

First thing you'll notice is how only the center can be focused, while from about one-third out to the edges the rest is pretty soft. Then up close you'll be able to see the gross color aberration: this will not only introduce rainbows toward the edges, but will soften the image closer to the center of the screen, as the color will not quite converge, thus causing an apparent loss of focus. It isn't. It's just that when the colors don't converge it appears to be due to a loss of focus. And appearances are everything in this game. In fact, appearances ARE the game, unfortunately.

Contrast will be shot, too. Much of your projector's light will reflect back into the projector optics, bounce around and cause ghosting as well. It's not so bad with AR coated lenses, but it'll still be pretty bad. And this is IF you get the axes of the lenses perfectly aligned. You will find that even a fraction of a degree in rotation and tilt is fatal to even a reasonably recognizable, much less a sharp image. Forget 1080p. Try for OK DVD quality at best. 300p YouTube standard might be a better bet.

Then look for soft spots in localized areas of the image. These will almost certainly occur because optical lenses are made to very loose specifications compared to those used for scientific applications as properly made anamorphic lenses are). Mostly they are moulded plastic, but even glass will exhibit local surface irregularities. The reason opticians get away with these slack standards is because the human eye can adapt to less-then-perfect focus, and because the human eye can only see a couple of degrees clearly, right in the center of the eyeball anyway. The rest is salesmanship. With projection you can inspect the entire screen for such flaws, at your leisure, of which there will be many in all likliehood.

For the first few days you'll be bowled over by having an anamorphic lens in your house. But as soon as you see even a half-way decent prismatic lens you'll sart to have doubts. When you see a high quality cylindrical lens, and what they can do plus what they DON'T do) to an image, you'' get the drift: there are no shortcuts to excellence in imaging, excellence enough to come anywhere near to matching your investment in an expensive projector and screen in the first place.

The anamorphic lens is the last thing the beam passes through on its way to the screen. Before it was installed in place, the projector lens was the last thing. Ask yourself why so much care is taken with projector lenses and why you think you can get away with so little care in its effective replacement. Perhaps then the penny will drop.

Last edited by Aussie Bob II; 03-08-2015 at 07:23 PM.
Aussie Bob II is offline  
post #57 of 60 Old 03-08-2015, 07:49 PM
Scott Horton, techht.com
 
GetGray's Avatar
 
Join Date: Aug 2004
Location: Mid-South USA
Posts: 5,569
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 93 Post(s)
Liked: 83
Quote:
Yeah that's a pretty rude elitist statement...
Not sure I ever remember being insulted here . That aside, I was trying to be neither. I was speaking to the OP and trying to save him some time, money and effort. Sorry if I popped your bubble and compelled you to chime in. As I told him, if he can do optical magic and defy some physics, I'm 100% behind him.
GetGray is offline  
post #58 of 60 Old 03-09-2015, 10:04 AM
Advanced Member
 
Metric's Avatar
 
Join Date: Jun 2001
Posts: 852
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 28 Post(s)
Liked: 17
Quote:
Originally Posted by Aussie Bob II View Post
It seems you won't be discouraged, so go ahead. Get Gray's statement was not elitist. It was a statement of fact.

Go ahead and try some eyeglass elements. You may even think you've discovered the Holy Grail. But unfortunately there is no el cheapo way of doing that. There are no dirty little secrets optical engineers keep to themselves.
No its not a statement of fact - The guy with the $1k projector doesn't need to stick with zoom. Perhaps in context to whats being asked here (cylindrical) but there are diy solutions to get an 'acceptable scope image'.

I want to die in my sleep like grandpa... not kicking and screaming like the people in his car.
Metric is offline  
post #59 of 60 Old 03-09-2015, 10:05 AM
Advanced Member
 
Metric's Avatar
 
Join Date: Jun 2001
Posts: 852
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 28 Post(s)
Liked: 17
Quote:
Originally Posted by GetGray View Post
Not sure I ever remember being insulted here . That aside, I was trying to be neither. I was speaking to the OP and trying to save him some time, money and effort. Sorry if I popped your bubble and compelled you to chime in. As I told him, if he can do optical magic and defy some physics, I'm 100% behind him.
I may have taken your comment as blanket when it was specific to the op's quest of cylindrical.

I want to die in my sleep like grandpa... not kicking and screaming like the people in his car.
Metric is offline  
post #60 of 60 Old 03-09-2015, 03:58 PM
Member
 
Aussie Bob II's Avatar
 
Join Date: Oct 2012
Location: Sydney, Australia
Posts: 76
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 17
Too much flame here.

You can't put a square peg in a round hole. That's the bottom line.
Aussie Bob II is offline  
Sponsored Links
Advertisement
 
Reply DIY Constant Height



Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off