Homemade Telecine

[Stan Jelavic]

Here is my first POC for a full custom telecine with the DSLR camera and my own transport. Used a new stepper with a timing wheel.
https://photos.app.goo.gl/yeB55FWBhQV4MFjP7
https://photos.app.goo.gl/ZgiGqW5B7sosLkNd8
https://photos.app.goo.gl/U76yLMkoGeHNmX71A
https://photos.app.goo.gl/QgWYhBeJPKwQQRUg9

The resulting video has a wobble, slowly drifting up and down.
https://photos.app.goo.gl/xeWD1PNgEoG9jzEe8

Believe that cause is the timing wheel tolerances. The notches have to be spaces very accurately because even a tiny variation becomes noticable.
Got the part done at Shapeways. The material used was Versatile Plastics which may not be the best choice. Perhaps Kamel, David or Bruce may have some suggestions here since I still do not have too much experience with 3D printing.

[Stan Jelavic]

https://photos.app.goo.gl/dHvhDfJV4ePiuGAAA

[Stan Jelavic]

Decided to give this idea one more chance. Will use fine detail plastic this time to get more precise timing. If it works then this could be used for a cheap but excellent quality homemade telecine. The film driving mechanism can be replaced by pinch rollers allowing for films with lots of missing perforations. Also it could be designed for any size film.
Another idea in case this does not work is to use the RPI project but just for frame timing.
https://www.youtube.com/watch?v=6vQSJZa7OGE&t=21s

The idea is to have two gates in series. The first gate has a RPI with the cheap RPI camera used just for timing of the stepper. Can have a cheap LED setup. The second gate is image capture gate. The stepper timing can be adjusted to compensate for the separation between the two gates. This is a bit more expensive but not that bad since the RPIs and RPI cameras are really low cost.

[Stuart Reid]

Looks interesting, I still have all the steppers from my previous aborted attempt to build a scanner, maybe I'll give this a go.

[Stan Jelavic]

Glad that there is some interest. If the timing disk approach works out then my main problem after that would be 3D printing. Still do not have the 3D printer and getting the stuff done outside gets pretty expensive.

[David Brown]

Hi Stan

Glad to see this is still moving forward! I've been out of the loop since early June. All my capture equip and film is in boxes and I hope to be up and running by the end of the month.

I do regret that I now have a lot of reading ahead of me, need to catch up.

[Stan Jelavic]

Welcome back David. I slowed down as well but still keep it going.

[David Brown]

Hey Stan looks like I have finished my reading, at least the new stuff. I will need to review things leading up to the "homemade telecine" and Bruce's struggle with HDR, though his issue is different than mine.

Captured video Looks pretty good!

Stability is very irregular and strange. It seems to pull down too much (>1.667 in.), then ramp down to less (<1.667 in.). The speed variations all cancel each other since we don't see frames continuously rolling down the screen.

https://drive.google.com/file/d/1p-K...ew?usp=sharing

I doubt the 3D print material type has any impact on accuracy, more likely that would be printer settings. Commercial prints should be as accurate as possible, or get your money back!

Are there visible imperfections on the 3D sprocket? Assuming this is an optical sensor, would a smaller "shutter" window help?


Using two notched wheels as a way to narrow the timing window (if necessary) without having to create multiple printed versions.

Does the optical sensor provide feed back to adjust stepper speed or trigger the camera?

Sorry, these questions are probably answered on earlier posts!

[Stan Jelavic]

The wheel notches synchronize the stepper. The stepper runs until there is a positive pulse from the wheel. Then it stops for a few seconds and sends the trigger to the cam and then continues running until it gets another positive pulse.
I can see imperfections in the print. They appear as small bumps. Going with fine detail plastics may be better. Ordered the part several day ago and should be getting it soon.
The width of the notches really does not matter since the MSP430 is detecting the first edge only.

[David Brown]

Hi Stan. Just playing cheerleader. You probably don't have parts yet or have doubts about accuracy of this method.

I think I will divert my interest from the DSLR project to this one. The Winait scanner I bought has instability in both x and y axes. Slowing the speed might help, not sure. Slowing it down only makes real time video capture even harder.

I still want a DSLR motion capture option. I can do still capture, it is a little cumbersome. So, looking for projector parts on ebay. Specifically gates, both 8 or S8. Either one could work if the framing mask is removable. Not sure if printing a gate is a good idea. Smooth surfaces are very important.

[Stan Jelavic]

Hi David. I did receive the new sync disk but it is too transparent so I painted it black. Will try it soon after finishing a few Hawkeye orders. People are still buying the kit but it is still small qty that does not justify auto assembly. So I do it manually which is a pretty tedious process. If this was my prime job I would be a poor man for sure, but I do it because I like it.
I agree, printing the new film gate is a pretty tricky business. You do not want to damage the precious old memories. The tension also has to be right etc etc.

[Stan Jelavic]

Here is a simple gate design David.
https://www.youtube.com/watch?v=ltHuyay-2Os
It does not touch the film only the edges which minimizes the film damage.

If done in smooth fine plastic it should not scratch the film.
Two rubber bands can be used to control the tension.
Alternatively, two springs with adjustable top bracket can be used.

Thought you may be interested in trying this.

[David Brown]

Thanks Stan, I just might test that simple gate idea. Odd, just 2 pieces of plastic over a light source, what's with the 4 wires??? !



I bought an 8mm carrier for my Cannon scanner several years ago, but I decided 8mm film was too small for the scanner's resolution. Flatbed scanners exaggerate their resolution, just like the Wolverine.

Several film carriers came with that scanner, but not 8mm. I bought one from Amazon (printed) made to fit my Canoscan 9000, about 6 years ago. It's still available for sale. It's not unlike the simple gate.



https://www.amazon.com/s?k=Negative+...ref=bl_dp_s_we b_0

Also, I found a sound projector selling for parts on Ebay. $40 and free shipping. Shipping has to cost ~$40? It will have a gate and there is a very large sprocket wheel that might be useful. What worries me is that it might be in working condition! The seller doesn't know if it works or not, it's missing a power cord!

[Stan Jelavic]

Hi David. Not sure why he has extra wires there. He has several designs and also managed to get the scan working with the stepper and Arduino which sounds promising.

[Stan Jelavic]

Tested the new sync wheel and it works better but the shake is still too high in my opinion.
https://drive.google.com/file/d/1sDx...ew?usp=sharing

Not sure what the cause of shake is. Perhaps the wheel tolerances or maybe the way the optical coupler triggers, not sure at this point.
Will have to rethink this approach I guess.
Here is the picture of the new rig.
https://photos.app.goo.gl/z3NQjxGjeUYZ9DsR8
Everything is pretty solid in the film transport so do not think that the shake is caused by the things being loose. Oh, well, back to the drawing board.