Film-Loop Monitor

Forum for discussion of narrow-bandwidth mechanical television

Moderators: Steve Anderson, Dave Moll, Andrew Davie

Film-Loop Monitor

Postby Andrew Davie » Wed Jan 24, 2007 2:56 pm

I have some interesting (for me, anyway :D ) ideas about alternate mechanical TV configurations (such as using a loop of 35mm film instead of a Nipkow disk, with the ‘holes’ being clear areas in the film, and passing this at fairly high speed over a LED array/monitor. This would of course allow a compact monitor – but more interesting to me, the use of film allows the overlapping of the ‘holes’ and the opaqueness of the overlapped edges to be set (say, 50% when overlapped) such that the lines between scanlines become “invisible”. Using 35mm film would presumably allow very precise positioning of the ‘holes’ (or transparent windows, really), because the film could be printed from computer generated output. This would also make it much cheaper, I assume, making a TV monitor. I’m aware that the film would need to be moved at pretty high speed.
Andrew Davie
 

Postby Klaas Robers » Wed Jan 24, 2007 2:56 pm

The idea of the film-loop monitor is very old. Drawings were published long ago. But there is a not so simple to see problem that prevents us from using it.

First the film has to pass the sprocket wheels so frequently that it will wear out very fast. Film and sprocket wheels are not designed to run so fast, it will be running 400 frames per second. Then the speeds and the forces are too large to be handled by the rather fragile plasic of the film.
However this can be overcome by glueing the loop in a circle onto a disc and producing a drum monitor.

But there is another problem that I will try to point out for you.
1. Suppose we have a system with 32 lines and 32 dots per line. In total this is 1000 dots per frame.
2. One frame has a transparent dot with a surface of 1/1000 of the total frame.
3. So 99,9% of the frame is black.
4. So we throw away 99,9% of the light en keep only 0,1%.
This explains why we need so much light and still remain a rather dim picture.
I use 32 high brightness LED's at 50 mA peak current. You can't look into that.

5. Each dot of the picture is illuminated by this light 0,1% of the time through the "hole" of the disc.
6. If you use film this is attenuated by the transparency of the hole, say 0,9.
7. So this gives an illumination of 0,09% of the brightness of the LED-array. Still about 0,1%.

8. But each dot of the picture is also illuminated in the remaining 99,9% of the time
9. This is attenuated by the darkness of the film.
10. Suppose the attenuation of the film is 500 x. This is already very good for silver halide film.
11. Then each dot is illuminated as well with 0,2% of the light of the array during the other dots.
12. Say that the light is on for 50% of the dots,
13. Then this stray-illumination is 0,1%, the same as the illumination through the "hole".
14. So white dots get an illumination of 0,2%,
15. and black dots an illumination of 0,1%.
16. The so called contrast ratio is 1 : 2, which is extremely poor.

I know that a contrast of 1:500 can be reached by Black and White (silver halide) film. There is also special graphical material that can do only total black and total white, no gray shades. I don't know its contrast ratio, may be it is somewhat better. But if we want to come to 1:100 we need an improvement of a factor of 50 x. I doubt this. Besides this material will only be available as sheets, not as 35mm film. There is no need for graphical material in the movie industry.......
Klaas Robers
 

Postby Stephen G. Mican » Wed Jan 24, 2007 2:57 pm

I have been interested in television systems of the optical scanning or “mechanical” type since about 1967 at the age of 17. I tinkered around with a film-loop based optical scanning system at that time, using a 16 mm motion picture camera recording a moving light to generate scanning apertures! The technique for generating the apertures on the film worked remarkably well, but the concept failed.

The problem with using photographic film as a potential optical scanning element is the opacity or contrast ratio of the film itself. Unfortunately, the amount of light that passes through the dark area of the film is the same or more than the light that can pass through the clear apertures recorded on the film. Possibly one could use the film as a template for etching a photo resist-covered thin metallic strip and then etch holes in the strip to use the strip as the scanning element, but I never went any further with that concept.

Although photographic film is not suitable for an optical scanning element due to its limited opacity, a CD or DVD substrate would be fine. I see that you are a computer programmer. If you could come up with an application to burn a CD or DVD with preferably square apertures arranged around the perimeter of the disk suitable for an optical scanning element, such an application would be a most valuable development.
Stephen G. Mican
 

Postby Jim Wood » Wed Jan 24, 2007 3:07 pm

Your idea for a film scanner goes back the heyday of mechanical TV; a rubberized-cloth belt in the original proposal. This has the decided advantages of a rectangular scanning area (as opposed to the Nipkov "wedge"), a flat-plane scanning area (contrasted with a drum scanner, which gives a rectangular scan but on a curved surface), and, as you say, a smaller viewer, as the film can be run back and forth over idlers. I believe that Doug Pitt tried this with either 16mm or maybe 8mm film, actually photographing a white card on which he could position a black square dot for each frame's exposure. Good luck with this, and please keep us posted of your progress.
Overlapping scanning holes does indeed reduce the obvious line structure without an appreciable degradation of resolution. Ovals and diamond apertures work well, too, and of course a rectangular aperture can 'just touch' on every line. Doug Pitt calculated the percentage overlap for round holes for me once (he's a former high-school Physics teacher and good with math(s) ), and it seems to me it was on the order of a 15% enlargement in hole size for equal brightness over the scanning area.
Jim Wood
 

Re: Film-Loop Monitor

Postby Andrew Davie » Wed Jan 24, 2007 3:11 pm

I didn't mention in my introductory email, but my idea for a 35mm film-based monitor can be extended to colour. Make the film 3 x longer, containing 3 repeats of the 'holes'. The first of the 3 has the holes coloured alternately red/green/blue, the next has them green/blue/red, the next blue/red/green, so each scanline is traversed by a red hole the first frame, followed by a green hole the next frame, followed by a blue hole the next frame... and repeat. Further, each triad of scanlines is scanned first by a rgb, then by a gbr, then by a brg. That is, one frame will consist of a red scanline, next to a green, next to a blue, next to a red. So each frame has a single colour per scanline, but the scanlines are alternately coloured... and this alternate colouring is changed on a per-frame basis. This is the system I used to get colour images into a B&W tv game system. It would be very interesting to see how it looked on a mechanical system (I suspect it might flicker a bit too much, alas).


Andrew Davie wrote:I have some interesting (for me, anyway :D ) ideas about alternate mechanical TV configurations (such as using a loop of 35mm film instead of a Nipkow disk, with the ‘holes’ being clear areas in the film, and passing this at fairly high speed over a LED array/monitor. This would of course allow a compact monitor – but more interesting to me, the use of film allows the overlapping of the ‘holes’ and the opaqueness of the overlapped edges to be set (say, 50% when overlapped) such that the lines between scanlines become “invisible”. Using 35mm film would presumably allow very precise positioning of the ‘holes’ (or transparent windows, really), because the film could be printed from computer generated output. This would also make it much cheaper, I assume, making a TV monitor. I’m aware that the film would need to be moved at pretty high speed.
Andrew Davie
 

Postby Guest » Wed Jan 24, 2007 3:12 pm

Hi Klass

I understand all your comments about the attenuation of film, and the low contrast ratio ensuing. I have an immediate comeback to those comments. Simply pasting TWO (or three!) of the films back to back would immediately significantly increase the contrast, no? If one wishes, instead of a clear area, the ‘holes’ could be actual holes in the film, introducing no additional attenuation in that area. This multiple film idea would also significantly strengthen the film itself for fast traversal through the mechanism.

One could also go to the effort of painting the film by hand, masking the black areas with some reflective silver paint, for example, further reducing attenuation and (?) also increasing the light available to the holes, if the LED array was backed by a mirror and the silver back to the film managed to reflect light back into the LED ‘box’.

Another idea I’m not so keen about is that the columns of the LED array could also be turned off when not needed (that is, say there were 4 columns of LEDs, then the first is only turned on when scanning the first (say) 10 or so scanlines, the second is turned on only when we scan lines (say) 8-18, etc – a bit of overlap there is intentional. I’m not so keen on this because I’m still rather uncomfortable using a LED system in the first place; it feels like cheating… so using the advantages a LED array has to offer is cheating even more. One could simulate this by having a secondary belt that ran perpendicular to the first, the belt only allowing light through for the scanline being drawn. This would reduce the light through the black area by 32x, of course.

To be honest, I’m not particularly interested in film wear – I have more interest in if such a system could be made to show a good picture (even if parts have to be replaced once a week, that’s OK!). Mind you, I’m nowhere near actually making this yet, this is just a ‘what if’ at the moment.

Cheers
A


Klaas Robers wrote:The idea of the film-loop monitor is very old. Drawings were published long ago. But there is a not so simple to see problem that prevents us from using it.

First the film has to pass the sprocket wheels so frequently that it will wear out very fast. Film and sprocket wheels are not designed to run so fast, it will be running 400 frames per second. Then the speeds and the forces are too large to be handled by the rather fragile plasic of the film.
However this can be overcome by glueing the loop in a circle onto a disc and producing a drum monitor.

But there is another problem that I will try to point out for you.
1. Suppose we have a system with 32 lines and 32 dots per line. In total this is 1000 dots per frame.
2. One frame has a transparent dot with a surface of 1/1000 of the total frame.
3. So 99,9% of the frame is black.
4. So we throw away 99,9% of the light en keep only 0,1%.
This explains why we need so much light and still remain a rather dim picture.
I use 32 high brightness LED's at 50 mA peak current. You can't look into that.

5. Each dot of the picture is illuminated by this light 0,1% of the time through the "hole" of the disc.
6. If you use film this is attenuated by the transparency of the hole, say 0,9.
7. So this gives an illumination of 0,09% of the brightness of the LED-array. Still about 0,1%.

8. But each dot of the picture is also illuminated in the remaining 99,9% of the time
9. This is attenuated by the darkness of the film.
10. Suppose the attenuation of the film is 500 x. This is already very good for silver halide film.
11. Then each dot is illuminated as well with 0,2% of the light of the array during the other dots.
12. Say that the light is on for 50% of the dots,
13. Then this stray-illumination is 0,1%, the same as the illumination through the "hole".
14. So white dots get an illumination of 0,2%,
15. and black dots an illumination of 0,1%.
16. The so called contrast ratio is 1 : 2, which is extremely poor.

I know that a contrast of 1:500 can be reached by Black and White (silver halide) film. There is also special graphical material that can do only total black and total white, no gray shades. I don't know its contrast ratio, may be it is somewhat better. But if we want to come to 1:100 we need an improvement of a factor of 50 x. I doubt this. Besides this material will only be available as sheets, not as 35mm film. There is no need for graphical material in the movie industry.......
Guest
 

Postby Klaas Robers » Wed Jan 24, 2007 8:38 pm

I fear that glueing two films back to back will obstruct their bendability far too much. When glueing them to a disc to form a drum might give problems with the difference in length that the two films should have, however you might make the apertures on one film somewhat larger, which also overcomes the problem of registration of both apertures. It is not needed to improve the contrast of the total area, if you do this for 99% of the area this makes the available contrast already much better.
You also could consider two film strips that run a different path and only come on top of each other in the viewing window. Then the worse bendability is overcome.
Klaas Robers
 

Postby Andrew Davie » Thu Jan 25, 2007 12:38 am

Klaas Robers wrote:You also could consider two film strips that run a different path and only come on top of each other in the viewing window. Then the worse bendability is overcome.


An interesting idea, too!

Not quite what you were suggesting (that is, two identical film strips with the same holes that coincide at the view area), here's another suggestion using the same thought of scanning two film strips...

Have two film strips that cross the viewing area perpendicular to each other. One is the normal Nipkow-type scanning holes, and the other is a 'scanline mask'. The scanline mask is such that wherever the other film's hole passes for a scanline (as the scanline mask is passing, too), the scanline mask is transparent. That is, it would be a diagonal on the film itself, and the speed of it would be such that the entire diagonal swiped the display area for a scanline in the time that single scanline was 'drawn'.

The whole idea is that the secondary film acts as a light filter, only allowing light to pass for the current scanline being drawn. The film would therefore consist of diagonal holes, each being the height of the display, and scanned one hole per scanline (horizontally). That is, there would be 32 scanline mask holes scanned per frame. THis system would attenuate the unwanted light to roughly 3% of what it would otherwise be (roughly 1/32 of the original unwanted light would still get through)
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1511
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Postby Klaas Robers » Fri Jan 26, 2007 12:28 am

I have a drawing of the film loop scanner from the thirties, but I don't know how to share this with you on the Forum...

Here we go...
Image
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1453
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Postby moderator » Fri Jan 26, 2007 12:46 am

Klaas Robers wrote:I have a drawing of the film loop scanner from the thirties, but I don't know how to share this with you on the Forum...


That would be very interesting! You can either link to an image on your own web space using the image tag as shown above the edit window when you're typing a message, or alternatively email me the image and I will be happy to enter it into your message for you.

I have also created a NBTV ftp area, so if you know how to use ftp, you can upload images to the web space I have provided, and then link to that.

ftp server: ftp.taswegian.com
user: nbtv
password: mechanical

When you upload files to the above, they will appear in the location http://www.taswegian.com/NBTV/images. So, for example, if you uploaded a file 'tv.jpg', you would link to it in an img tag using the URL http://www.taswegian.com/NBTV/images/tv.jpg

Following is an example...

Image




Cheers
A
User avatar
moderator
Site Admin
 
Posts: 112
Joined: Wed Jan 24, 2007 2:40 pm


Return to Mechanical NBTV

Who is online

Users browsing this forum: No registered users and 5 guests