LD-50 (Lethal Dose - 50%, I'm led to understand) was shot 35mm on Kodak's new Vision2 film stock, a stock ideally suited to the DI process due to its extended dynamic range, enhanced linear response and reduced grain.
The film was produced by Random Harvest/Four Horsemen Films, directed by Simon De Selva and shot by DoP Robin Vidgeon, BSC and the DI performed at the Moving Picture Company [MPC] in Soho, London via their iQ/Qcolour based DI suite.
MPC's iQ DI suite
Offline was performed traditionally via Avid and the resulting edl used to 2K scan the selected takes via a Spirit DataCine to generate over length scans.
The scanned selected takes were autoconformed, checked against the offline video version, and grading commenced with client involvement throughout, with Director, Editor, Producer and DoP often being in the room at the same time.
This proved to be one of the major benefits of the DI approach - enabling the client to interest with the grading process in a way impossible with traditional film grading.
During grading, shots requiring vfx work were networked into effects systems such as Shake & Inferno. Although the DI system (iQ) is capable of performing such work, business sense suggests using cheaper parallel systems while maintaining high throughput within the DI 'hub' system. The resulting vfx shots were networked back into the DI system for inclusion within the edit, replacing 'shot holding' offline version, and then graded in context.
Within the film were shots captured via VideoCam and these had to be 'merged' with the 35mm footage, while retaining a user pov feel. The two main challenges here were to mix 2 different resolutions (2K and SD) and colour spaces (Linear TV and Log film).
This project worked almost exactly as a DI project should. the edl was accurate, and although no reel was 'locked' until the last moment the changes were easy to track and the online kept accurate.
The negative was provided as flash-to-flash cut selects reels making scanning relatively straightforward and the scanning 'pull-list' generated from the edl worked almost 100% accurately.
The autoconform from edl worked very well and matched to the offline video just about 100% for each reel.
The client was able to interact with the project throughout and gain a result that was closer to their original vision than otherwise possible.
Almost a perfect DI project from beginning to end with the whole operation taking about 3 weeks to complete.
After the 'film' was output via the ArriLaser it was necessary to provide video deliverables too. HD and SD, text-less and with text, all generated from the DI data so identical to the 'film' output in every way.
No audio work was performed within the DI suite, although the offline audio was used to check sync during the grading process.
|