DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Techniques for Independent Production (https://www.dvinfo.net/forum/techniques-independent-production/)
-   -   Mini 35 Competing Unit for a Lot Less $ (https://www.dvinfo.net/forum/techniques-independent-production/10532-mini-35-competing-unit-lot-less.html)

Kai Leibrandt June 29th, 2003 10:29 AM

Sorry to reply to my own posting, but I thought I'd just give a quick summary of what I had achieved during my _quick_ experiments.
What I tried to do was to create a really simple, cheap (less than $5.00) proof of concept and experiment that would show me the various things that were involved in creating a homebuilt mini35.
I have an old, broken Olympus OM-2 from which I took the lenses, a wide angle 35-105 Tokina zoom and a 70-210 tele. I also have an Angenieux 12-120 16mm lens lying around that I put through the same treatment...
I used a bit of cardboard tubing as the envelope holding target, for I used a CD-case cover (the square front bit that is), This piece I initially sanded to make it opaque, but the sanding did leave streaks, so I used a piece of tracing paper instead, which has a very coarse grain but at least it was even.
The results were surprisingly good. I have just pointed the XL1 with my manual 16x at the "ground glass" and got some ok pictures, but nowhere near good enough to use of coarse. I used macro so that the lens was actually only an inch or so away from the target, and of course this would not allow the image to fill the viewfinder completely, but it did show some gorgeous depth of field, and also, surprisingly, the colours were gorgeous too. It was also very interesting to see what exactly it means when people say there 16mm lens will not cover a 35mm image; the diameter of the image the Angenieux projected onto the target was less than half of what the other lenses produced. At the same focal length however, items in view would be the exact same size on the image, only you'd see less the surrounding objects around it with the 16mm lens.
Also clearly visible was the vignetting that I was expecting as I don't have any fresnel lens or similar material. What I found out is that the ground glass of the OM-2 is in fact ground on a curved surface which points away from the lens, and the other side is flat (and shiny). I beleive this is to compensate for the light fall-off toward the edge of the viewfinder, and I will try to do something similar with my setup.
I will probably post some pictures sometime during the week, and until then I hope to be going around the local camera repair shop to hunt for some old lenses and bayonets to make something sturdier, plus I hope to find a ground glass similar to the OM-2's so that I don't have to take my camera apart, even though it is broken...


Cosmin Rotaru June 30th, 2003 02:15 AM

just one side

Kai Leibrandt June 30th, 2003 05:15 AM


yes I did discover that also. I did wind up taking my OM-2 apart and that little bit of ground glass is a real piece of magic. One side has a fresnel on it, so fine you can't even see it, you can only feel it if you (gently!) go over it with a finger nail. It's this frensle that makes it look curved while it is in fact perfectly flat. The other side (which points away from the lens) has the matte finish to it. Results were very promising and I am getting some more bits today to start building a more solid one.
Will report back when more interesting stuff happened...


Cosmin Rotaru July 1st, 2003 03:00 AM

Hi again,
Jaime, here is what I understand about the fresnel:

In this first pic I try to explain what happens to the light when it hits the GG. Of course you'll have an image forming on the GG, the light spreading out. But the light tends to follow its original path, so a part of the light that hits a point on the GG would spread out but most of it will go straight ahead, as it hit the GG.
The fresnel would concentrate the light before it hits the GG (or after - you could chouse to install the fresnel on the other side).

I hope this helps!

What else have you guys try?
Joseph, have you found some other means to vibrate the GG?

Jaime Roman July 1st, 2003 03:26 PM

I will now be incorporating a fresnel lens in my configuration -- thank you Cosmin :)

Now all i need to iron out is my source for vibrating the GG.

Joseph George July 3rd, 2003 01:45 PM

I got bunch of private emails -- I was out of middle of nowhere and except for email did not have real Internet access. Thanks. What I was going to send Cosmin was the info on ground glass. Here it is for everyone. Both the Mini and the Pro units have circularb GG's. It looks like a filter and it is in a filter-like barrel. Edmunds Scientific sells ground glass. They are on the web. Get the finest GG you can get.

John Jay July 4th, 2003 03:24 PM


it looks to me that (in after effects terms) the gg is displacement mapping the original image at the film plane (especially your non vibrating gg cuddly toy pic)

i wonder if a physical engineering solution is the most cost effective way ahead?

maybe if the gg signature was known (as a greyscale image) an inverse displacement map could be achieved in after effects? so the gg distortion could be illiminated in post.

an experiment to determine gg signature would have to be devised of course (not easy but a once off solution)

it makes me think that the superb Takumar lenses M42 thread could be used, screwed on to an M42 fixed length extension tube with the gg contained inside the tube, then screwed via step ring to virtually any cam with macro function and less than 42 thread

Cosmin Rotaru July 7th, 2003 01:17 AM

Hmm... The inverse displacement map sounds good in theory. I don't know how to do it, though... Could you use a pic for the displacement map?

Elmar Tewes July 7th, 2003 03:02 PM

i finally found something that looks really interesting...
it seems there is another company who brings out his version of a dv - 35mm system - perhaps a lot cheaper than the p&s technik one ???

Jaime Roman July 7th, 2003 10:15 PM

i found a few more pics of movietube online at this site:


Only problem is, how does it work?

Joseph George July 8th, 2003 12:14 AM

Very interesting. Does anyone speak German to translate this stuff?

Cosmin Rotaru July 8th, 2003 02:25 AM

very interesting, indeed!

I tried an online transaltion (nothing interesting):
"The Movietube is an accessorie for miniDV cameras in order to simulate the Videodrehen the Tiefenunschärfe of real cinema films. The Flash animation on the CD-ROMS generated for a patent competition is actual component of the Movietube homepages that goes soon Online. For the competition, the presentation in Flash was modified for the use as an alone standing program on a CD something. Soon there will be the total presentation and to see more Online. Beforehand here schonmal a couple Screenshots of the CD (to the Vergößern please simply on the picture click)"

Jaime Roman July 8th, 2003 07:43 AM

This is the translation i got for the beginning of that statement

" The Movietube is accessories for mini digital video cameras, in order to simulate with video tricks the depth and sharpness of genuine motion picture films."

John Jay July 8th, 2003 11:25 AM


I have two proposals to consider one is software related and the other is physical, i shall start with software

taking your clear cuddly toys pic (straight from video) and displacement mapping it with a greyscale noise pic in photoshop would produce an image quite similar to the stationary gg pic of the cuddly toys, provided the displacement factor was set to a suitable value.

So in effect, the displacement of pixels of the clear pic is proportional to the intensity of the grey value of the pixels in the noise pic. This process is reversible (try it) since taking the displaced pic from above and displacement mapping it again but in the opposite direction (negative dispacement factor) would produce the original clear picture.

ok so far?

what is required is to generate a greyscale noise pic which corresponds to the effect of the stationary gg. Some experimentation is required. The following is only a suggestion.

A good start would be to generate some fine black vertical lines on a sheet of white paper (thickness of the lines =space between them) and to measure how much they are horizontally displaced after passing through your stationary gg rig (I expect some zooming in will be required). The result of this will deliver the horizontal noise component of the gg

Next repeat the above with fine black horizontal lines to deliver the vertical noise component of the gg

Then composite the horizontal and vertical components into a single greyscale noise pic which in theory should be the noise signature of the stationary ground glass

bear in mind the resolution of the noise pic should be one pixel, which is tedious i know , so it may be best to concentrate on a small area to check that it works in practice.

my second proposal is a possible physical solution

Consider that your 35mm lens (from your rig) ultimately will be attached to a fixed length extension tube. Now with a circular saw cut a slit in the extension tube perpendicular to its axis at the point where the filmplane should be but not all the way through.

Make a circular disc gg of say CD proportions - a CD gg!

Spin the CD gg at high speed (5000 rpm is possible) through the slit created above by means of a motor mounted on the outside of the extension tube, this should sort the distortion problem and reduce the need for high quality ground glass. (maybe running the CD gg at slower speeds might introduce some desirable grainy effects)

Maybe the CD need not be ground glass at all but just radial score marks if it spins fast enough

A final design would have a housing for the CD gg and motor and would look a bit like a 'machine gun kelly' type of rig if you have understood me correctly

Cosmin Rotaru July 9th, 2003 01:47 AM

Thanks for the ideas, John.

Actualy, the second one I already tried a few days ago. I have a small (about 1.5MB) clip that I could send it to anyone interested. But only tomorow because I don't have it at me (sorry Elmar, I forgot...). The image is better than with the vibrating stuff.

The first idea is somnething that I have to test. I didn't relize is so complicated! The first time you mention it I thought I'd just take a pic at a white paper to see the grain... I din't think of the vertical and horizontal lines.

I'll let you know how it works.

John Jay July 9th, 2003 07:03 AM

please send it


Jaime Roman July 9th, 2003 10:23 AM

Hey Cosmin,

Please send me that clip also to -- romanemp@aol.com.

Are you also a member of the homebuilt stabilizers forum? I believe i just finished answering a question you had on aluminum alloys.

Cosmin Rotaru July 9th, 2003 10:53 AM

Thank you Jaime! Yes, that's me! :)

Spencer Houck July 11th, 2003 09:51 AM

Could you send the clip to me also?

And, sorry if you have said this, what camera are you using? Seems everyones working on Xl1 35 adapters, but i'm "stuck" (terrible thing to say about a purchase as major as this one was for me) with my Sony VX2000, and would love to see some headway in that department.

Thanks for the clip and inspiration. I'd love to get back to Ohio University next year and blow some ppl away with any ressemblence of shallow DoF. :)


Cosmin Rotaru July 11th, 2003 10:47 AM

I've sent you the clip, Spencer. I've sent the clip to anyone who ask for it!
I already received some feedback on my e-mail. Thanks! It seems that the image is nice, so I'll continue experimenting. I'm not going to try eluding the existing patents and make something for sale. I just want to make something for me... But I'll help anyone who whants to build their own.
I own an Canon XM2 (that's the European version of the GL2).

I like very much how the movietube looks! The adapter looks nicer than the P+S one. With shoulder support, viewfinder, XLR adapters... A complete accesory for miniDV.

I expect prices going down from now, as there's competition. I realy don't understand how could they ask 8000USD for an accesory made for a 3000USD camcorder...

Spencer Houck July 11th, 2003 12:12 PM

I love that image, absolutely great!

I'm a very uninformed "noob" if you will, but I keep hearing people say that the 35 mill conversion doesn't work well with cameras with permanent lenses, such as your GL2 and my VX2000. Your tests seem to put that thought to shame, and I hope the same can be attained by my VX noting our similar lens setups.

Knowing that I'm totally uninformed, in theory could we in effect cancel out that we have lenses in front of our cameras? Bear with me now, but could you replicate the permanent lens structure of our cams in reverse and toss it in front of our cameras to take us down to a sort of "glassless state" like the Xl1? It seems a little fantastical and may introduce color abberations, etc., or maybe it just doesn't make any logical sense at all, but it just sounds like it may be crazy enough to work to someone who's clueless :). Who knows, maybe with your setup the glass in front of my camera won't instigate a problem, I'll just have to test and find out.

Oh yeah, I can't see very well what the philatelic lens(5x) looks like in your setup picture. Can you explain that to me, is it just a normal magnifying glass like any one you'd find?

Thanks for keeping the innovative stuff alive,
Spencer Houck

John Jay July 11th, 2003 02:08 PM

eureka or maybe countach!


1 was the fresnel still necessary?

2 looks like I shall find a new use for the lenses from my Pentax LX kit :)

Jaime Roman July 11th, 2003 10:09 PM

Hey cosmin,

How did you invert your image? Did you do this in post or did you use prisms?

I took a canon eos camera apart and am using a skeletal version of it; problem is i didn't realize that canon eos lenses do not have a diagphram control it is controlled by the camera body (which i just finished destroying, UGH!) do these lenses default to being wide open? Or, will i be destroying a perfectly good lens?

Cosmin Rotaru July 14th, 2003 03:34 AM

Hi guys!

Spencer, why do you need that "glassless state" for your VX2000? Al you need is to focus the camcorder lenses onto the 35mm GG. For this you need a close up lens: a normal magnifying glass.

John, in this second test I didn't use a fresnell lens. The image is cropped on top and bottom (for a 16:9 or whatever look) and so the corners are left out. And so is the vigneting!

Jaime, I inverted the image in post. I don't have a prism. I don't know about the EOS lens. The lenses I use (Pentax K-mount?!) are wide open by default. The lens has a small lever that goes from the lens in the camera. Whatever aperture you chose, the lens stays open until you take the pic. Than the aperture is closed at the chosen level by switching that lever on the lens (yes, the camera does that). You can turn that lever by hand and see how the iris closes. I don't know if this is the case with the EOS lenses.

Spencer Houck July 14th, 2003 06:56 AM

Could you post some pictures of your newly updated rig, especially any with the camera set up on it also?

I seem to have a hard time focusing my camera on anything relatively close which is discouraging me a lot in recreating this great effect. I believe the GL2's minimum focusing distance is much shorter than my VX2000's. Hopefully I can get it all working.

Spencer Houck

Cosmin Rotaru July 14th, 2003 08:11 AM

I'll see what I can do so I can take a pic "with the camera set up on" the rig.
The problem is that I use the camcorder to take the pics...
You'll see that the GG is not very close to the camcorder's lens. Look again at the pics I've already posted. Is the same setup but I use the rotating CD instead of the GG/fresnel sandwich. The magnifying glass allows zooming in the GG.

All times are GMT -6. The time now is 11:08 PM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2021 The Digital Video Information Network