RED Image Plane Splitter for 3d Imaging / Z-Matte Creation / Dual Long-Shot&Wide-Shot
I don't know what the total deal with this is since it's similar to other ideas (but a bit more in-depth), but here's my idea:
Create a camera lens mount that accepts 2 lenses next to each other. The lenses would be geared so as to allow changes in interocular distance while maintaining focus. It would also incorporate a dual-lens follow focus system (one follow focus wheel turns both focus rings) with a dual-aperature lens control system as well (perhaps a smaller wheel).
These can obviously be decoupled so attaching two different focal length lenses allows you to create both a wide shot and closeup from the 1 camera during the same take. There would have to be some adjustment per-lens/mirror system that would allow locking backfocus between the lenses (for stereo imaging) or seperating it (for the wide/tele shooting system).
The images from both lenses hit mirrors prior to entering the sensor area, and are designed to create dual 1920x1080 images next to each other (seperated by a few pixels, at least, to prevent lighth leaking and aid in setting up recording areas) on the sensor (perhaps an internal matte box on the mirrors to reduce the image size on the light coming in from each lens to 1920x1080).
Then a recording mode that will output dual 1920x1080 image streams from the 2 images produced on the sensor. If there's slight overscan from the mirrors, maybe in software you could pick what two areas of the chip you want to record the streams from, or just record the overscanned images so you can match them up precisely in post.
Then, there would be an add-on binocular-style viewfinder device, so the operator can determine the proper interocular distance for the shot. A breakout box that outputs a 48-fps compatible stereo image (for shutter glasses) OR the old color seperation (for those old color glasses or instances where you can't use the stereo monitors for framing) would be included as well.
Per-pixel tracking software (like PFZed) could determine image planes for matte extraction from the dual image streams.
I don't know what the total deal with this is since it's similar to other ideas (but a bit more in-depth), but here's my idea:
Create a camera lens mount that accepts 2 lenses next to each other. The lenses would be geared so as to allow changes in interocular distance while maintaining focus. It would also incorporate a dual-lens follow focus system (one follow focus wheel turns both focus rings) with a dual-aperature lens control system as well (perhaps a smaller wheel).
These can obviously be decoupled so attaching two different focal length lenses allows you to create both a wide shot and closeup from the 1 camera during the same take. There would have to be some adjustment per-lens/mirror system that would allow locking backfocus between the lenses (for stereo imaging) or seperating it (for the wide/tele shooting system).
The images from both lenses hit mirrors prior to entering the sensor area, and are designed to create dual 1920x1080 images next to each other (seperated by a few pixels, at least, to prevent lighth leaking and aid in setting up recording areas) on the sensor (perhaps an internal matte box on the mirrors to reduce the image size on the light coming in from each lens to 1920x1080).
Then a recording mode that will output dual 1920x1080 image streams from the 2 images produced on the sensor. If there's slight overscan from the mirrors, maybe in software you could pick what two areas of the chip you want to record the streams from, or just record the overscanned images so you can match them up precisely in post.
Then, there would be an add-on binocular-style viewfinder device, so the operator can determine the proper interocular distance for the shot. A breakout box that outputs a 48-fps compatible stereo image (for shutter glasses) OR the old color seperation (for those old color glasses or instances where you can't use the stereo monitors for framing) would be included as well.
Per-pixel tracking software (like PFZed) could determine image planes for matte extraction from the dual image streams.
Last edited: