A Rumblehouse Telecine System Design Project

Real-time 3D printed system objective

 

Introduction: Where am I on this project. What is yet to be done?

This is phase two of my telecine system update for HomeDVD. The first system was built about 8 years ago. It does and continues to serve me well, but it is time for an upgrade. HomeDVD was the only transfer shop in Canada offering true HD film telecine services at the time.

The first system imaging core was a Lumenera based 2K HD camera system, running on a heavily modified Moviestuff projection system and controlled using a custom written Windows based software application.

Now I want to look at developing a new telecine system that runs faster than 6-8 frames per second, can capture at better than 2K frame sizes, can handle different film gauges from the same platform and squeeze more out of the dynamic range offered by the camera system – that is: deeper blacks, smoother uncompressed whites, better shade discrimination and to capture any details hidden in the darks and in the highlights of a film frame.

In addition to these goals, I want to share my in-progress development work with fellow telecine DIY'ers and 3D print enthusiasts.

Come on in and join the conversation. There will be comment fields below every information page.

In a previous blog post, I have documented that small gauge film just does not have the line pair resolution for better than 1920 pixels – particularly when degradation effects due to aging, questionable lab post processes at that time and the average quality of the original film stock all takings its toll. The digitized films are already over sampled at 1920, just based on the embedded lumpy grain sizes small gauge film inherently presents. I still maintain this point, but lets see if other performance factors can be improved with better cameras, lenses and post processes.

My first pass at this project is to use my newly adopted 3D CAD and print skills to design and build custom made parts coupled with off the shelf (OTS) hardware that is readily available. As it turns out a lot of the OTS parts I plan to use are available from Amazon. So sign me up for Prime.

After reviewing many of the possible parts and assemblies found in the main stream DIY printer videos and forums, a design emerged that got me very interested in pursuing my next HD telecine project. The key points being that the critical custom parts are to be 3D printed and the balance of parts will be sourced off the shelf. My main goal being to have an updated in-house system followed by a productized system that can be easily built and replicated by 3D print and DIY telecine enthusiasts.

I got a good mix of ideas by reading forums and reviewing lots of Youtube videos showing other DIY telecine efforts out there. Lots of off the wall stuff for sure – yet amazingly workable. What was curious was that many DIY’ers used parts from existing yet dated projectors and other home made parts to make their project a reality – cool stuff but a non starter for me. It’s tough making a copy of the DIY project if you can’t get the parts required. My design had to be reasonably doable and have no restricted access to reasonably priced parts.

My main parts supplier as it turned out is Amazon, with odds and sods from other suppliers. Of the multitude of products available at Amazon, a lot of what I need (and to keep my budget in check) are parts sourced from China – of course, which for the most part presents the issue of lengthy delivery times hanging over your head, even though the shipping costs are either surprisingly low or free. You want cheap, you gotta wait. You want it yesterday, crack open your wallet and depart with more cash.

Other than a reliable and available source of OTS parts, the 3D printed parts required had to be relatively easy to design and simple to print using standard print filaments. The 3D printed assemblies are to be of high quality, be precise in operation, rugged and be reliable under continuous operation. The film pulleys must all have bearings, the tension arms to have springs, the main sprocket wheel is to be precision made and camera positioning assembly is to be a 3 axis stage concept. The entire film path must ensure the film does not waver, get misguided, keeps tension (no jerky pulls and pushes by unbalanced film reels) and not ride up the roller guides and pulleys while the system is running.

My main concern is precision and consistent timing in film frame tracking. My preliminary design to address this concern has been built and tested and looks very promising. The big nut to crack from my point of view was to address the means to advance the film as fast as possible (ideally in real-time) and still capture rock solid well registered images – repeatably and reliably frame to frame. That is to ensure - to capture a well registered film frame, apply de-mosaic and convert the film frame to a 4:4:4 bitmap in the planned frame resolution and within the time limit before the next frame grab.

If this key framing and registration concept did not work, I would have to fall back on other ideas. That may mean longer development cycles, more hardware and connectivity considerations and perhaps incur a performance loss as a compromise.

All of my 3D printed parts will be available to purchase either as a real printed part or as an STL file so you can print you own. A complete BOM (Bill of Materials) and an assembly guide will also be made available.

I’m in the very beginning stages of my actual design and build so there is still a lot to think about. To help to organize my thoughts I have developed a 'to do' list that I want to share. I would certainly encourage feedback while the project progresses as I will post lots of info as the project progresses.

Use the comment field at the bottom of this and any of my linked info pages to register your thoughts, questions and feedback.

Planned info pages to be posted - Not in any particular order

Linked descriptions, pictures and videos highlighting system build progress.

Current posting date - Jan 2019
Running update Log.....
  • put Date: - sample entry field -brief what was done, changes etc

  • Designspark Mechanical – Free 3D CAD software – design all of my custom system parts
  • 3D Printer tools and materials – Simplify 3D slicer, PLA and PETG filaments, printer selected
  • Steppers and sprocket wheels – design, math, considerations, print material
  • The Arduino – sketches, prototyping and test control functions
  • The cage – main parts mounting platform, OTS {link here}
  • Mounting hardware – selected screws, washers, T-slots, rack joiners, Metric only by choice
  • The film path build plate – rollers, film gate, drive sprocket (pinch rollers), film path tension
  • Camera 2 axis stage control plates – X and Y movement
  • Camera Z gantry and camera support – lead screw movement assembly, camera attachment
  • The film gate – exact width and guide, back light support, single plate design
  • Hi-energy LED back light, heat sink, dimming control, diffusion, vignette effect
  • Film guides and rollers – bearing inserts, spring tension loaded, slotted guidance, the film path
  • Tension arms and springs – DIY spring jigs, control arm pivots
  • Tension sensor controlled feedback for smooth feed forward play
  • Take-up reel design, assembly and control – slip pulleys, belts, motors, hi-speed fast forward - TBD
  • Source reel design, assembly and control – slip pulleys, belts, motors, hi-speed rewind - TBD
  • Reel spindles – spring loaded locking clutch for 8, S8 and 16mm reel pins
  • System control – switches, etc
  • Custom electronics – Xtal clocks and timing chain, triggers, phase locked loops, power supplies
  • Selecting a camera – sensor size, SDK, global shutter, external trigger, C-mount, capture methods
  • Selecting a lens – C-mount, focal length, glass quality, line pairs
  • Custom software application development – Windows *.NET, camera SDK
  • Application computer hosting considerations – camera interfaces, port speeds, user interface

Leave a Comment