Homemade Telecine

[David Brown]

Oh no, not the drawing board! Hope you enjoy puzzles.

The frame movement does not seem to have a repeating rhythm. There are 243 frames in the video.

The bottom of the frame never drops below the starting point of the first frame.

Up and down movements are small with occasional large jumps.

Frame 36 hit the highest point above starting, once and never again.

I don't know how much work it would be to bypass the wheel and light sensor.

Have the motor move a fixed amount (0.1667"), wait for the camera, and repeat?

Seems easy, but I don't know how to do that.

[Stan Jelavic]

Sorry David, posted a filtered video.
Here is the raw video with the sync wheel:
https://drive.google.com/file/d/12-m...ew?usp=sharing

And here is the raw video without the sync wheel, just a fixed motor advance frame by frame.
https://drive.google.com/file/d/1cRB...ew?usp=sharing

The second video still has a drift which is not a big concern, but the shake is still there. It could be caused by the film stretch. I believe you had a similar issue. I will fix the drift tomorrow and then rerun the test. Will also compare the same footage with the Hawkeye capture.

Actually, I like the drawing boards and puzzles, keeps my brain going.


Here is the raw Hawkeye clip. There is some shake but seems to be lot less than in the second video. Will have to put them side by side.
https://drive.google.com/file/d/1qM2...ew?usp=sharing

[David Brown]

Stan, I wish you the best here. I'm going to cry UNCLE.

Unless it is possible for the motor to move different distances, I can't understand the moving frame line. If the step distance moves more or less than the film frame spacing, the frame line would continuously drift up or down.

Sadly, using the motor alone looks worse than it does with the sensor wheel.

The raw Hawkeye clip looks ok to me. There are tiny frame line movements up and down. If the image is cropped that is not perceptible with the camera panning.

[Stan Jelavic]

It does drift but it has a jitter (shake) around it. The drift can be eliminated by tuning the number of steps per frame. I got pretty close but still needs more work. The sync wheel can also be used as a feedback to get the drift eliminated (phase locked loop). But the main problem is the shake. I will check the stepper tomorrow by pointing the camera at the sync wheel and observe the edges for shake.
In this project below a stepper was also used and the test clip also has some shake similar to mine.
https://www.instructables.com/28-mm-Digital-Telecine/
Also
https://www.chrisgavin.com/search/label/telecine
So, looks like the shake is inherited in the transport using the stepper motors. Gris Gavin used the After Effects stabilizer and the resulting video seems pretty good.

[Stan Jelavic]

Actually this setup may work. With the gate that I am using the perforations are not visible. I will try a different cover with a larger opening to get the perforations in the camera view. That way a de-shaker can be used. The sync wheel still may be needed to eliminate the drift since the freewheeling method without the sync wheel will always have some very small drift preset.
OK opened up the gate opening and now I can see the sprocket hole:
https://photos.app.goo.gl/KfUAiZeuHPJf69n36
Will try the de-shake avi script from Mattias:
https://8mmforum.film-tech.com/vbb/f...3022#post13022

[Stan Jelavic]

Playing with the de-shake script. The script posted by Mattias has some syntax errors do to forum web site mangling it. Mostly spaces added. But the bigger problem are the additional plugins needed.

LoadPlugin("plugins/masktools2.dll")
LoadPlugin("plugins/variableblur.dll")

These plugins are not a part of standard Avisynth or VideoFred releases.
See plugin summary page:
http://avisynth.nl/index.php/Externa...Download_sites

Or go directly here.
http://avisynth.nl/index.php/MaskTools2
http://avisynth.nl/index.php/VariableBlur

Note that gausianblur will not run on all window machines. A windows DLL is needed (use 32 bit version):
See notes here:
http://avisynth.nl/index.php/VariableBlur

[David Brown]

The instructables 28mm telecine video has a lot of instability. I noticed it in X & Y axis and a little rotation too. I've seen all of that when loosening the gate tension of the Wolverine or the Winait machines. When gate tension is really loose even the framing can change.

Stan, since you're using a projector gate you should not have the stability problems of a DIY. That's if the motor has plenty of torque, and the takeup is not pulling too hard.

I believe by shake you mean the bouncing frame line in the x axis?

I've decided to not tightly crop new captures for two reasons. Sometimes there can be too much movement in X and then I won't need to recapture the whole clip. The second reason is to use frame and perforation for software stabilizing, when needed.

I remember Mattias had all versions of AVIsynth in both 32 and 64bit loaded and was able to manage as needed. I'm doing everything in Vegas now. I could not load any of your sample avis from post #17 into Vegas or Davinci. They play in the browser or with VLC. I was able to load them into Vdub.

So Vdub2 has Deshaker(possibly others) and gausianblur filters. I think they come stock, if that makes it easier. Offhand I don't know why he would want to blur to anything, I'm always trying to remove it! Unsharp mask is my goto sharpener.

[Stan Jelavic]

Hi David, with the good quality gate the shake is only in the direction of film travel. I ran the deshaker that Mattias posted and it does a very good job and the result is a stable video but I get some black frames in the video. Will check that tomorrow.
https://drive.google.com/file/d/1uUc...ew?usp=sharing

Got it working.
https://drive.google.com/file/d/1j14...ew?usp=sharing
Disabled the mode line from the script.
#SetMTMode(2,2)
But believe the main culprit is my laptop which sometimes is very slow at times.

[David Brown]

That's looking good! I should stabilize everything. There is no loss of image quality, just a little loss of real estate around the edges.

And even if the film capture is stable, there can be an unstable camera man!

What are the black frames? Are they extra or missing frames?

[Stan Jelavic]

Exactly, just some loss of the area around the edges. After trimming it should look good. There is a hole under the gate for illumination. I had piece of white paper there that is supposed to reflect the light up onto the gate. The paper got ripped up during installation and that could be causing uneven light distribution. Will look into that later. As you can see the led mount is pretty primitive and I did not expect anything big from it.
The black frames are caused by some of the original frames changed to black. Believe that is caused by multithreading. So it really depends on how fast your machine it. Multithreading is faster but could cause some issues such as blank (black) frames. That is my guess.

[Stan Jelavic]

Fixed the LED issue.
https://photos.app.goo.gl/yQA9EnRodFFh8zXP9
The hole drilled into the wood mounting plate was not wide enough and not centered. The reflections from the side of the hole were very bright and causing the bloom. If the hole is wider than the bottom gate then the reflections have no way of reaching the film.
I am pretty amazed how a simple concept like this can produce good quality images.
Here is the LED assembly sketch:

[David Brown]

I didn't notice the led light leak, but I can clearly see it now that you described it. Could have been glare from the water, nice catch!

Is is possible to slow the capture motor down when the computer falls behind for whatever reason?

I just received the projector I bought. Shipping was free, now I know why. Three weeks in transit. The last 10 days within 10 miles of me??? Fedex handed off to the Post Office. Maybe they were practicing for Friday's drop in delivery speeds?

Can't complain about slow for free. I'm about to begin to dismantle the gate and maybe the sprocket wheels. I'll be sure and go back thru your posts for the assembly of the gate and light.

Quality images from a quality source and camera!

[Stan Jelavic]

Free is always good. The speed of the motor can be adjusted as slow as you need it. The gate, the camera and uniform light are all key components to good quality. It should work ok with lots of types of cameras with some type of macro lens (or reverse lens) since there are no mechanical restrictions.

[Stan Jelavic]

Here is the captured video run through deshaker.
https://drive.google.com/file/d/1kmm...ew?usp=sharing

[David Brown]

Well done Stan!

The stabilizer seems to take care both issues then, drifting and shakiness.

Do you plan to move to a capstan/pinch roller from the sprocket gear now?

I was going to remove the large sprocket and the gate from the projector I just bought, but then there's always regular 8mm in the wings.

A capstan could do both. That also means the motor's increment will need to change for each format.

But then, this projector I bought for parts, is working! A friend with a sound projector has offered to loan it to me in the future, though I don't know it's condition. I do know my new projector works and it has a speed control for fine tuning sound.

The gate is a little odd and might be difficult to mount outside of the projector's frame.

I guess I'm going back to EBAY! I must keep this new projector running.

I do have some sound films that I have put off capturing until the silent ones are finished. That could happen before this project is finished.

Using a sprocket will be more difficult to include regular 8mm. Only one gate is needed, but two different sprockets.

That also means a different increment for the stepper motor.

If I jump right to a pinch roller, no need for a reg 8mm sprocket!

These demo films you just posted look better. Most of it seems to be the color, and they could have more detail. Just a little, not crazy better.