Wolverine-Hawkeye Telecine

[Stan Jelavic]

Don't disagree Brian. But doing it from scratch is a pretty big undertaking, bigger than what we are doing here IMHO.
If someone in the forum wants to start it however I will support the project.
Possibly easier task would be just to take the guts out and create your own mounting hardware. David is doing that.

But anyways eyes back on the ball. Pulsing the takeup from the external 12V supply works OK. The takeup barely budges if the pulse is short. The motor current is 250mA so I can just use a 2N4401 transistor and a protection diode and that is it.
No, no problem with the existing boards David. Will do both. An external small board and the V12. Good point however.

[David Brown]

C'est magnifique ! Kamel


The set screws went in like butter. Just tighten each side a little, adjust the pulley support, tighten again and BAM! Locked and straight.

[Chris LaRocque]

A question of the group; I've read through the forum and seen Jitter mentioned several times. From context I believe this is a vertical alignment repeat-ability issue. Can one of you comment on the scale of the issue and the current remediation procedure?

Thanks

Chris

[David Brown]

Chris

This guy is calling this jitter https://youtu.be/Vy7CBI0iiQM I would agree!


I have not had this problem. I only digitized about 15 minutes of film when I decided to FIX the quality.

The cause of jitter has been described as
1. A slow memory card. It would need to be several years old.
2. 8mm that was badly split from the 16mm width. Leaving at least 50% too wide.
3. An aggressive takeup reel that constantly pulls on the film, interfering with the claw.
4. Excessive gate pressure on the film preventing the claw from advancing a complete frame in one move.

I could hear the sound of the claw fighting the takeup, and physically restrained the takeup reel. Not the best method for all the movies I have! Maybe that's why I have not had any jitter in my 15 minutes of captured film.

Stan has overcome most if not all the Wolverine's faults. This last pass he has taken care of the takeup by controlling the takeup motor. That motor is independent of the claw motor and though there is a slip clutch, the Wolverine designers eventually added all the rubber coated posts to introduce more friction in the film transport.

gate pressure can be reduced, slip clutch tension can be reduced (not your first choice), buy a faster memory card, takeup speed reduced.

Maybe I've missed something?

[Stan Jelavic]

David covered most of it. Some jitter is present in the gate itself or worn film but the biggest offender is the takeup that tugs on the film at the end of the claw cycle causing misalignment at the beginning of the claw cycle when the picture is taken. I have done changes to mitigate the takeup caused jitter by pulsing the takeup. The pulsing did not work well because the pulse was routed through the Wolverine power board and that one has a minimum pulse width constraint. That minimum pulse was too long.
So today I moved the pulse circuit onto the Hawkeye controller and can have a very fine pulse control.
The limited tests that I ran today at different frame rates look very encouraging. Almost no jitter.
Here is the video of the gate operation at 2FPS. Also have a capstan running.
https://photos.app.goo.gl/Cz3PQHfB18j4PKfd7

Here is the capture with IC Capture. Very little jitter:
https://photos.app.goo.gl/TM2qS6wYP3kZXhVK7

[Kamel Ikhlef]

Thank you David, glad the new pulley works.
Stan, Jitter was really a big problem for me and I'll be very happy if you can fix it.

[David Brown]

Stan I saw 3 frame jumps. The bottom frame line is in view, then drops below view, returns and repeats 2 more times.
Looks like torn perforations would act. I can't see the whole perforation to be sure if that is true.
As it drops, there is motion blur, that is most likely the video camera or the film is moving during capture.

I have had the Wolverine lose framing after recording started. That couldn't be the claw, it has no adjustment or way to shift, right?.
It had to be the the image crop position on the sensor jumped. That however, did not return to the original position.

If that drop happens on the same frames each time recorded, there is something odd with the film.
I saw nothing else in that entire clip that I would consider a problem.

Kamel does this jitter happen to your captures when there is no perforation damage to the film? Even if there is no take up, where the film falls into a box?

[Stan Jelavic]

I know that this film had some problems at the beginning David. Will look at the tiffs and then rerun to see what the issue is.
Ok, figured it out. The film perforations in this clip have some issues. Reran the video without takeup and slow speed and got the same jitter.
Original problem frame
https://photos.app.goo.gl/mM4rzWBLtg5vZDFX8

The same frame without takeup and 0.5 FPS:
https://photos.app.goo.gl/mM4rzWBLtg5vZDFX8

There is no blur in the tiff images so that must be my cellphone cam. Would like to add that some small jitter can be caused by the gate pressure and inserting a folded piece of paper between the gate and the cover can help as documented on the Wolverine web site.

New takeup circuit. Can be done as a small daughter board.

[Chris LaRocque]

Thanks Stan and David for the clips. That aligns with my expectations and brief experience. This is actually my second pass with the Wolverine. My first was with the 720P (cheaper) unit a couple of years ago which I returned after 24 hours. I watched the Pro for quite some time before deciding to purchase again. I was going to send it back as well but discovered this forum and decided to invest in the improved version. I'm working on a python opencv capture kernel that will crop the frame reliably. My plan is to use the full resolution of the camera with a slightly wider FOV exposing the frame edges as a reference for the crop. As soon as I get the detection core right I'll share.

Stan, when can I send money for the v11 board?

C.

[Stan Jelavic]

Don't worry for money now Chris. I am getting ready to start assembling the boards. It will be V12 with the new takeup connector.


Will have to order the boards and still short of some parts. Opencv and python is a very powerful solution. I have been using it for over 4 years now at work for photogrammetry.

[Stan Jelavic]

Hi Kamel,
So it looks like you are pretty happy with the new back pulley and the new toothed pulley that you shared with us. Any plans for the regular 8?

[Kamel Ikhlef]

Hello Stan ...
I have no R8 film to test, but I will design according to your tests and measurements of the pulleys
corresponding to the R8 and post them as soon as finished. (maybe tomorrow night).

I will take this opportunity to redo the other pulleys with the updated dimensions ...
Thank you a lot for the V12 update (with all the improvements).
I will finally opt for the V12 with the new takeup control which seems very good to improve the jitter.

David, the regular jitter I have is random with any movie (it's really a wolverine flaw - a lot of testimony on internet).
With this jitter problem, I started to despair about this telecine ! But, Stan...

... Stan is doing us a great service in solving this big problem and I can't wait to try v12 Hawkeye.

[David Brown]

Chris and Stan may I call you Photogrameters?

Ok, my understanding is that is used for making 3D files from 2D scans. Drones capturing terrain and cell phones scanning parts.

This sounds Chris that you are planning to crop the captured image via software? My first thought is can we lose the claw? Don't know. There is a slight amount of movement inherent with film (jitter) and it is safer to crop in tight, projectors mask off a smaller image than the camera recorded.

My second thought is, can the software crop and align the images? That would eliminate any jitter whether from damaged perforations or random jitter like Kamel is describing.

This is exciting!

I have been capturing today with a new 16mm lens. I can't evaluate that now, but I have seen some frame jumps that might be what Kamel is describing, but I was also making speed changes without stopping so that might have been the result of my meddling. I will know more after some time. It does look good.

Stan, I'm more than happy to have V11 and a daughter board. After all, who knows what the final version will be?

[Stan Jelavic]

Thank you Kamel. I do have some normal 8 footage but only a few.
Yes David, that is it. Generating point clouds from 2D images and processing them. Calculating object dims etc.
For the daughter board I would have to move the MSP onto it to intercept pin3. And then it will be a few connections from the daughter board to the main board. It is probably Ok if you already have a board and want to add the takeup cct to it. But if you are getting a whole new board I suggest you go with the V12. It can still operate in the V11 mode. The only difference is MSP formware. But if you want to go V11 then V11 it is I have qty 4 V11 blank boards, and one fully assembled.

Kamel and David, is the jitter with the takeup running? What type of pulleys are used? Is the takeup pulsed or continuous?

[Bruce Davis]

Hi all, there is much happening, I ordered a v11 board but now we are moving to v12, I can not keep up. Stan please change my order to v12.

I have seen considerable improvement in jitter with the last MSP430 that Stan programmed for me (although seem to have lost continuous takeup, but never used it anyway), the take up reel timing was changed and doubling of capture speed was really great. Previously it was difficult to minimise the capture frame size because of the considerable jitter.

Have always been able to eliminate jitter with post processing.

Chris it is great to hear that you are working on being able to crop images accurately and automatically using the frame borders. I have in the past, looked on forums for such an option but found nothing.

Have made alterations to the Capstan pulley parts, these are meant to be printed with a 0.15mm layer height, or if you wish 0.075mm layer height.

Pain full with the small file size upload limit, so link below:-
https://1drv.ms/u/s!AtIUdJCJJJGieEc-...dlNCE?e=0gEDKe
The measurements should be as follows :-
S8 Capstan Pulley Assy

N8 Capstan Pulley Assy.