Webcam Video Assist

[carllooper]

Carl, Expanding the vertical axis to fit the 16mm frame height would also expand the horizontal axis resulting in image loss. The x and y axis cannot be uncoupled in standard optics.

There's always going to be some loss during an optical print but ignoring that, I want to expand the frame height without expanding the width. Doing this means there is less loss in the vertical than there is in the horizontal because there is more of the print film being used to capture the vertical signal than there is to capture the horizontal signal. Even better would be to do the same but to 35mm. But sticking with 16mm for the time being.

That's the problem I want to solve, how to do it optically rather than digitally: uncoupling the x and y axis as you put it.

Certainly this can't be done using a spherical lens but it can be done using some sort of an anamorphic lens, although I'm not sure what kind of anamorphic lens would be required, other than it must be a 2X lens and operate at the magnifications one uses for optical printing.

The anamorphic lens is necessary if I want to solve the stated problem. Without such I've been considering the following optical print setup:

The source is on a projector, with a conventional 2X anamorphic lens on the projector (but rotated 90 degrees from normal use) projecting a 4:3 version of the UP8 source:

1. On to a wall (or some suitable surface), which is rephotographed by a 4:3 16mm camera with conventional lens (quasi optical print)
2. Through a condensor lens, onto the same 4:3 16mm camera.

The result is the problem solved in terms of the sort after geometry (4:3 geometry), although both introduce more losses than the desired macro anamorphic. But I'm happy to live with either. Using the condensor lens will give a better result than the wall scenario. Using the wall scenario means I don't need to find a condensor lens. Originally I hadn't considered either. I thought it was a macro anamorphic I'd need, or it would be impossible. So I was experimenting with lenses to construct a macro anamorphic but couldn't (well not yet anyway). And then the wall idea occured, followed quickly by the condensor lens idea. Both of which delight me, especially since both are relatively achievable within a short time frame. Not the best but better than no solution at all.

If your were able to utilize an inline anamorphic projector lens you are essentially magnifying (expanding) the original UP8 camera grain. Which may be an intentional effect on your part.

Yes, the original UP8 grain is stretched (expanded vertically) but this is only on the print. This vertically stretched grain (on the print) is unstretched during screening, because the screening uses a normal 2X anamorphic lens (cinemascope lens) which unstreteches it. Or to put it another way the horizontally squeezed grain is expanded to match the vertical grain (treated as normal). The original UP8 image (and it's original grain structure) is restored on the screen. The audience see the intended result: the UltraPan8 image, but projected on a 16mm projector with a cinemascope lens on the projector. What will be horizontally stretched (or vertically squeezed depending on your point of view), will be the grain of the print film, but the structure of the original UP8 image and it's grain will be normal - and it will be more visible than the print grain (especially if I use a fine grain print film).

I have seen auctions on occasion with rare custom optical printer anamorphic optics. But they are quite rare and usually custom apochromatics.

Yes, I've been looking all over the shop. Where does one normally buy such a thing?

Carl

ps. Obviously there isn't any of this problem going the digital intermediate route - and that option is always there - but the problem being posed here was how to do a film to film print (for a film screening), without doing an digital intermediate.

pps. check this out: https://www.thingiverse.com/thing:81153

[Nicholas Kovats]

Carl,

That is a neat anamorphic prism link by the way. Did you ever invest in a 3D printer?

Well I shifted my perspective and clarity ensued. For some reason I was fixed on the projector side of things. You can possibly achieve your intended effect by experimenting with the following, i.e.

1. SOURCE: Your projector is the primary unaltered UP8 image source.
2. INPUT: Attach either a 16mm or 35mm anamorphic projection optic onto the taking camera prime lens. They are very common on eBay. I have a superb Bell&Howell 16mm specimen if you need to experiment. I also have a very unique monster "variable" anamorphic projection lens for the 35mm format. I believe It varies from 1:33 to 1:2.66.
3. X/Y ROTATION: More specifically attach the anamorphic projection optic to a carrier such that you could rotate the anamorphic oval horizontally, i.e. expand the x axis and contract the y axis. This is reverse of the traditional method.
4. OUTPUT: Set up the anamorphic optic in the same manner on the projector for viewing. This will reverse the input method, i.e. contract the x axis and expand the y axis.

I am still having a hard time to visualize the end effect. The original UP8 design intent was to remove the need for expansion/compression of x/y axis. For example if a conventionally oriented 2x anamorphic attachement were to be used in conjunction with a Bolex UP8 2.8 camera the resultant uncompressed aspect ratio created is 1:5.x or greater.

Check out this link regarding anamorphic prism testing and the software used to generate the images, i.e.
http://kb-en.radiantzemax.com/Print50227.aspx
https://www.radiantzemax.com/demo_13

[carllooper]

I am still having a hard time to visualize the end effect. The original UP8 design intent was to remove the need for expansion/compression of x/y axis. For example if a conventionally oriented 2x anamorphic attachement were to be used in conjunction with a Bolex UP8 2.8 camera the resultant uncompressed aspect ratio created is 1:5.x or greater.

Yes, I don't know if we're quite on the same page yet - but almost. In case the intended workflow (and resulting effect) is still confusing, see the diagram at the end.

So I want to maintain the UP8 philosophy. Shooting UP8 saves mucking around with anamorphic lenses. And is also cost effective. The first feature perhaps being more important than the second.

But the UP8 philosophy doesn't say anything about how to then screen the film, ie. for an audience? A UP8 projector is obviously the purist solution, and building a UP8 projector might be a future project. A digital delivery/intermediate is another solution, and this will tend to be the norm (and it's a setup I've already built). And a third option is a very traditional option: and that is to create a 16mm optical print made from the UP8 camera original, but using an anamorphic lens during printing. This is done in the studio (obviously) but it's where working with an anamorphic lens need not be so painful. I don't know that this necessarily goes against the UP8 philosophy. I figure that what it might lose in terms of UP8's utilitarian optics philosophy it might gain in terms of photochemical film philosophy. Now in this case an anamorphic lens is then also used (required) during screening of the print. This is a very traditional way of screening. And it's the method I've used in the past when doing cinemascope on 35mm. The image is squashed on the print and unsquashed on the screen. But it does require the projection venue have a cinemascope lens for the projector. However that's slightly more achievable than requiring the venue have a UP8 projector.

The desired effect is not anything strange: the end effect on the screen is just a normal UP8 aspect ratio image.

Now I already have an anamorphic lens (a cinemascope lens). That is not the issue. The lens is one I picked up off ebay for a reasonable price (a Bausch and Lomb) and it's a lens normally used for screening 16mm films in which a cinemascope image is otherwise printed on the 16mm film, in 4:3 format (ie. squashed), and the lens is then used to unsquash it during screening. And that's how I'll be using the lens.

What I don't have is an anamorphic lens for squashing the image on to the print film in the first place. The one I've got (for screening) might have worked but I haven't been able to get it to work at the close magnifications (macro photography) required for optical printing (from UP8 to 16mm). However I can get it to work if I project the UP8 squashed onto a wall (and rephotograph that with normal optics on 16mm). Or I could get it to work if I use a condensor lens - although the condensor lens would have to be quite large - however old school overhead projectors can be mined for such large glass (if I recall).

[milesandjules]

Hi Carl
Wouldn't you need a pan8 projector to project on the wall in the first place…if so why can't you just use that for projection for the audience? Im sure I've missed something here though.
Cheers
Miles

[carllooper]

Hi Carl
Wouldn't you need a pan8 projector to project on the wall in the first place…if so why can't you just use that for projection for the audience? Im sure I've missed something here though.
Cheers
Miles

That's a very good point.

The simple answer is that I don't have a UP8 projector. But the question remains: how do I then project it on a wall?

The nature of UP8 is such that you can run it on a 16mm projector. All the sprockets are the same as 16mm sprockets, and every second sprocket mates with the 16mm sprocket wheel. Projecting UP8 with a 16mm projector displays two UP8 frames - one above the other. And I have a 16mm projector.

But wouldn't this have to be converted to what is effectively a UP8 projector for the printing to work? Yes and no.

For the wall scenario I would use a mirror to switch between one frame and the other. I can do this in a frame by frame printing setup, but doing this live for an audience, at 24 fps, I couldn't. Well, not without using a different method such as build a UP8 projector.

It may be that building a UP8 projector is what I end up eventually doing - after testing (in theory or practice) each of the easier options.

C

[Nicholas Kovats]

Carl,

I appreciate the additional clarification and your minimalist sketch!

The unspoken issue is one of geography. You reside halfway around the world! I would lend you my UP8 2.8 projector in a heartbeat if you were local. I actually have three of my S8/16mm camera systems out with local filmmakers. I am a proponent of active usage of film equipment as opposed to dust collection. Well, that's not entirely accurate. There are still way too many gorgeous camera dust bunnies sitting on my shelves.

Back to the discussion.

The proposed anamorphic must be on the print camera not the projector. Think of it this way. Light rays entering the front of the anamorphic optic are compressed (oval: X-/Y+) whilst light rays entering the rear are uncompressed (oval: X+/Y-). The resultant projected print and aspect ratio would exceed the UP8 spec. You are essentially compressing the source UP8 frame and expanding the horizontal significantly.

You are describing the lack of close focus on a typical anamorphic 16mm projector optic. They are spec'd minimally from 5 to 7+ feet from the front element of the lens.

A 35mm projector anamorphic might be more suitable for your experiments as it will have a greater magnification effect at the center of the lens relative to it's 16mm cousin. There is a flood of beautiful 35mm projector integrated anamorphics from ISCO (German) on the market, e.g. http://tinyurl.com/mpubcqe. ISCO has been been consistent over many decades and they speak anamorphic. Mind you your working distances will be significant.

I wonder if the distortion would be tolerable if you were to use a 16mm or 35mm anamorphic optic in conjunction with a reversed prime lens on the printing camera? This is a very old method to achieve macro photography, e.g. http://www.cs.mtu.edu/~shene/DigiCam/Us ... ersal.html. You could potentially utilize your 3D printer to create a prime lens filter ring to C-Mount adapter to to attach the prime lens to the Bolex.

But truth be told alot of optical printer optics utilized the incredibly sharp Nikon macro lenses called Nikkor Micros and they were produced in various focal lengths, e.g. 50, 55, 105, 200mm. These would significantly reduce your working distance. They are very high resolution, minimal distortion and very good build quality, i.e. http://www.mir.com.my/rb/photography/co ... cronikkor/

In essence I am starting to describe a JK OPtical printer system with a slight modification for your setup, i.e. Bolex (print cam) + bellows on track + Nikkor Micro lens + 35mm non-integrated Anamorphic optic. You would have to experiment with the optimal working distance between the Nikon and the anamorphic optic potentially with a second bellows or helicoid adapter. The first bellows would be used to manipulate the magnification. I am in regular contact with a gentleman who has a very similar setup (minus anamorphic) and I can put you in touch.

Some examples of 35mm non-integrated Anamorphic optics aka adapters, i.e.

Schneider (2x) - covers 4 perf 35mm academy format
http://tinyurl.com/qgujarc
http://tinyurl.com/ols2mhq

Letus AnamorphX Adapter (1.33) - new - covers 3 perf Super 35mm format
http://letusdirect.com/cart/letus-anamo ... apter.html

Ok. Back to work.

Nicholas

[carllooper]

Hi Nicholas,

for the wall printing scenario it doesn't really matter whether the anamorphic lens is on the projector, or on the camera. The goal is to obtain a horizontally compressed image (on the print). If the projector is given the anamorphic lens, the projected image on the wall will be 4:3 (horizontally compressed) and a normal (spherical) lens on the camera just acts to transcribe this compressed image, as is, to the print film. If the anamorphic lens is used on the camera, it will be the projector which has the normal lens: ie. the projected image on the wall will be 8:3 (uncompressed) and it will be the anamorphic lens on the camera that does the horizontal compression.

Either way the image ends up horizontally compressed on the print film. As desired. Whether that first takes place on the wall, or on the film, doesn't really matter. The end result is the same. Six of one and half a dozen of the other.

From a more practical point of view the anamorphic lens I've got doesn't actually work on the camera as it would need to be closer to the film than the Bolex prism allows.

For clarification (but you already know this) an anamorphic lens can be used to horizontally compress an image or horizontally expand an image depending on it's rotational orientation - this is why you can use anamorphics originally meant for projectors on a camera and vice versa (well some better than others). Because you can just rotate the lens by 90 degrees to obtain the requisite effect. So if the task is to compress an 8:3 image to a 4:3 image on the projector you would not use the anamorphic as it is normally meant to be used on a projector, otherwise you'd get a 16:3 image! You would rotate it by 90 degrees in order to get a 4:3 image.

Now importantly it doesn't matter whether you consider this operation as the horizontal being compressed (and the vertical staying the same), or the vertical being expanded (and the horizontal staying the same). Both ways of describing the transform mean exactly the same thing. By way of explanation:

Speaking of a 4:3 ratio is exactly the same thing as speaking of an 8:6 ratio. They are the same thing. You will agree?

So if we were to speak of transforming an 8:3 image to an 8:6 image (using an anamorphic lens), we'd be saying exactly the same thing as saying we're transforming an 8:3 image to a 4:3 image, ie. we'd be using the same lens in the same way.

The only difference is semantic where we might describe one set of numbers as expanding the vertical (from 3 to 6) and the other as compressing the horizontal (from 8 to 4). But they are mathematically equivalent operations and more importantly (lest this be ignored) they have the same physical meaning. So there is no difference between saying compressing the horizontal, or expanding the vertical. They are equivalent ways of speaking about the same physical operation. Or to put it another, the orientation of the lens is exactly the same whichever of these two legitimate descriptions you use.

But there are two different ways of using the lens, and they will be describable this way:

1. Compressing the horizontal (same thing as saying: expanding the vertical) OR
2. Compressing the vertical (same thing as saying: expanding the horizontal)

For printing, I'm using the first case. And for screening to an audience I'm using the second case.

Re. a normal optical printer setup - I've already built one, with appropriate lens (or adequate!), lens/bellows adapter (3D printed), bellows, pan/tilt/roll/crane/dolly mounts on camera, stepper motor on the bolex, computer control of the projector - and we've made some great prints. It's not much different from the setup used to do film to digital transfers except that the Bolex is used where a digital camera was otherwise used.

C

[doug]

Hi Carl, Regarding rephotography of a projected image have you had any decent results from these possiblities:

1. using a long focal length projection lens, eg from slide proj, so the image is small at a reasonable distance from the camera (allowing use of anamorphic on projector). Tele lens on camera.
2. fitting a good quality closeup lens to the front of the anamorphic, allowing a closer projection throw.
3. perhaps taking the anamorphic apart and adjusting the focus range :roll:

I often do rephotography myself onto bits of card, and I've found with a Kowa 8Z anamorphic and close-up lens the results can be pretty good, though I try to avoid it if possible. (Come to think of it, I haven't done this for a long time. Usually I would put the CU lens on the anamorphic when fitted to a camera.)

To control the increased contrast I usually flash the image, either in a separate pass or using ambient light in the room.
Apologies for above if you already know all about this...
result of insomnia last night !

The idea of converting UP8 to a manageable film format such as 16mm Scope sounds good

[carllooper]

Hi Doug,

the anamorphic lens I've been experimenting with is actually a composite lens, meant to replace a normal projector lens rather than adapt an existing one. It's rated focal length is 2.5 inches (63.5 mm) and it projects the film onto a wall just fine. I haven't done any rephotography yet - just tested the setup by eye. I still have to go out and shoot some camera original. The flash idea is a good one. I've inadvertently flashed exposure in some other experiments, from light filtering through a 3D printed lens adapter (printed in white plastic).

Looking at the lens again I see the rear element can be removed fairly easily (has some set screws), so might try some experiments using just the front anamorphic component with some other optics.

A UP8 to 16mm Scope conversion. That's a good way of putting it.

The alternative to optical printing is by means of a digital intermediate and that setup is working fine. Much easier to do in many ways. The squeeze is done in the digital domain before a print back to film. And I can do a film out from such to 16mm or 35mm. For 35mm I'd get someone else to do it. But for 16mm I get a chance to do it myself (and control the process more). I'm waiting on the arrival of a 4K digital screen for printing digital intermediates back to 16mm. That should be here next week. It was the cheapest 4K screen I could find (about $400 new) but it's had mixed reviews in terms of it's colour gamut and luminosity fidelity - but those shortcomings are actually quite irrelevant when it comes to film out transfers. One just generates an image which takes into account the screens shortcomings. The cool thing about film is that you can do long exposures where you modulate the screen image during the exposure to get what you need on the film. For example, you could have an image on the screen where the pixels were constrained to being either fully on (white) or fully off (black). During exposure software just switches each pixel on and off, at any number of rates, to get a grey scale signal, of any bit depth you want. Indeed, if using panchromatic film the screen becomes a 12K screen (where each colour component of a pixel becomes individually addressable as a pixel). And for colour (as much as black and white) it's really just a matter of generating pixels (and exposure times) that correct for the screen's misallignment in terms of various colour standards. Colour is a bit of a bug bear. Ideally one would want a monochomatic screen (with independantly controllable filter wheel in front). Not just for colour film but orthochromatic film. In practice some experimentation with lens focus can premix (blur) the rgb components of the screen pixel.

But a digital pipeline takes something away from film. What it gains in ease of use and certain image quality criteria it loses in terms of other image quality criteria. The film to film pipeline (whether optical or contact) preserves the dynamic resolution of the original (contact printing better than optical), albeit at the expense of some increase in grain (in the original as much as the print). In a film to digital pipeline there's an increase in noise (hwever small) but more importantly there's an inherent limit on the high frequencies that can be transferred, and the result is a flatter signal, and it doesn't matter how much fake grain you add - or indeed how much real grain you add (eg. the grain of the print film in a film out). The signal is flatter than it otherwise might be. In the film to digital pipeline I've been testing special sampling techniques and algorithms to improve this flatness (at both a hardware and software level) and this is an ongoing project. The film to film approach is just an alternative, but with it's own specifically beautiful and spooky qualities that continue to challenge (for me) anything I would or could otherwise ever do in the digital domain. Its a quality that is really hard to ignore and I wouldn't want to ignore.

By the same token there is stuff one can do in the digital domain that is impossible in the analog domain. Its not spooky as such (or not as spooky) but it's special and specific in it's own way.

cheers
Carl

[milesandjules]

Hi Carl
Here is a crazy Idea…..could you rig up your pan 8 camera to project?

[carllooper]

Hi Carl
Here is a crazy Idea…..could you rig up your pan 8 camera to project?

Ha ha - yes that's a great idea - would require making a replacement pressure plate for the Bolex, with an appropriate aperture in it. But that's not too much work (considering the alternatives) and it would work just fine. And a little mirror behind it - to funnel the light through.

For projecting to an audience it's not as good as a conventional projector because the shutter blades are not divided up in the same way, ie. would exhibit more flicker than a conventional projector - but for transfer work it'd be just fine.

Great idea.

cheers
Carl

[MovieStuff]

I guess my question is why bother with anamorphics on the 16mm print at all? Just optically print the width of the Pan8 image to the width of the 16mm film and use spherical lenses to project. For maximum resolution you could even locate the image edge to edge between the sprocket holes and just leave black top and bottom. You'd have to enlarge the gate but that's easy. Or, for use on a standard 16mm projector without an enlarged gate, just print the width of the Pan8 image to fit the normal width of a 16mm image. If you did that, and had a two reel movie, you could put part one at the top of the 16mm frame and part two at the bottom upside down. No need to rewind! Just mask off the bottom of the frame during projection. When you get to the end, trade reel positions and run the second half!

Roger

[Will2]

I guess my question is why bother with anamorphics on the 16mm print at all?

Because that's what we discuss here; going the extra mile even when the first mile is probably just fine.

Am I right in assuming it would be the same as shooting standard 16mm, cropping to cinemascope then blowing through an anamorphic lens to the full standard 16mm frame (like in the awesome smiley drawing)? I would wonder what the real advantage of that would be too... maybe it would matter more on a larger screen and more powerful projector.

[MovieStuff]

I guess my question is why bother with anamorphics on the 16mm print at all?

Because that's what we discuss here; going the extra mile even when the first mile is probably just fine.

Am I right in assuming it would be the same as shooting standard 16mm, cropping to cinemascope then blowing through an anamorphic lens to the full standard 16mm frame (like in the awesome smiley drawing)? I would wonder what the real advantage of that would be too... maybe it would matter more on a larger screen and more powerful projector.

Well, that's kind of my point, in a round about way. I have to imagine that the image quality of the Pan8 is terrific from what I've seen but it's ultimately kind of a solution looking for a problem, and I mean that with great respect to the engineering involved, because you could do exactly what I'm suggesting which is to shoot 16mm with the top and bottom masked to get the same effect. However, the advantage of shooting in Pan8 and then printing to the top and bottom of the 16mm frame is that you can cut your print costs in half because a single print would carry both parts and one and two of the same movie, presuming that there are parts one and two, or even two different movies on the same roll.

But if maximum image quality is the key, then printing to the edges of the film and positioning the image between the top and bottom sprocket holes, you at least are dealing with the simplicity of spherical lenses, both for optical printing and projection. That, alone, should maintain as much of the original sharpness of the Pan8 frame as anything. Obviously, using the entire height of the 16mm frame anamorphically would be better but the question is whether it's really worth it in the long run. Spherical optical printing is easy and enlarging the gate of a 16mm projector is easy, too.

Roger

[carllooper]

I don't have any real reason for using an anamorphic. In many ways one will get a better print not using an anamorphic, at least in terms of the anamorphic options being considered. The use of the anamorphic is actually just a hangover from having designed the digital intermediate route for anamorphic projection.

The work I have in mind will cut between optically printed UP8 and digitally mastered UP8 but that could just as easily occur without anamorphic projection as much as with. But I'm banking on the possibility that even when film is rephotographed off a wall (handicapped in that way) that it will have a certain competitive look in relation to the digitally mediated image. Not that such is any reason for short changing the film. It's a bit of a quandry really. I want the digital image on it's best foot and printing for scope projection would do that. But I also want the film on it's best foot ...

One of the really difficult issues, whether using an anamorphic or not, is the switching between the upper and lower frame. The mirror setup on this needs to be finely tuned otherwise there can be oscillations in geometry between every frame. Using the UP8 camera as a printing projector is a simple solution and one I really like.

I've just ordered some LEDs and a controller for the printing projector - have been using a desklamp (and a daylight balanced studio light) but LED control will be a lot better. Have been using some white paper as a diffuser screen. But for the LED array this won't be quite sufficient. The separate RGB lights need to be merged into a single uniform even distribution of light. I've come up with a nifty solution here - two sheets of lenticular screen - at right angles to each other, placed between the LED array and the film does the trick. The result is a perfectly uniform and colour mixed backlight for the film. And bright - one stop better than using paper as a diffuser and one stop less than if no diffuser was used. It needs to stand back a little from the film plane to complete the mix. Optically the lenticular screen is creating a large array of tiny images, each of which is a complete image of the entire LED array. So any variation one makes in the LED array where it might be biased this way or that (in the LED array) gets redistributed across the lenticular screen - maintaining a uniform colour mix. And it's quite efficient in terms of light output. And cool to look at under the microscope.

Just noticed I've edited this post 12 times (this being the 13th). That notification usually happens if someone has responded after your post, (and if you are having an argument it's a bit of a cheat to rewrite the past) but in this case it's because I was editing this post for the first time when the previous post was submitted. I tend to re-read what I've written after I've posted it, and changed my mind on what or how I'm saying something ...
C