Lead Editor

Pre-Planning:

Team of editors (four, including lead) were assembled about 1-2 weeks before read-through. Initial priorities included workflow and skill determination/practice. Although each editor had prior experience in a different editing software each, it was decided to conduct this project consistently in Adobe Premiere Pro with the MIT-provided Adobe Creative Cloud license. Initial pre-planning meeting established a workflow spreadsheet to track scene progress [https://docs.google.com/spreadsheets/d/1-dte72LYIXhkHDk-rguAZS7BL0yS111kzVoSeNMe5ko/edit] and assigned full download/installation of Premiere to test system compatibility (computational power, disk space, etc.) and skills (keying, sequence construction, etc.).

After establishing the visual style with the director, a 3rd-person filming setup was devised. The primary goal was to make eye-lines automatic for actors, and embed natural blocking as much as possible into the technical setup. Actors were decided to act live opposite one another, with footage and audio captured from a separate device and location. Budget was allocated to individual ring lights/stands (for personal smartphones to film), microphones, and bluetooth headphones (as necessary). In addition to technology, green screens were acquired in the form of two green, plastic picnic tablecloths per person. Each actor was required to set up their space in accordance with guidelines provided by the lead editor [https://docs.google.com/document/d/1HprIne7kUcNQjxbm-5Vgov6B2-xuzyngmtPF0XrUQJs/edit]. Spaces were checked through live Zoom meetings one-on-one for each actor to help work through spatial limitations and answer any questions about the technical requirements. Individual test footage was requested for approval prior to the initiation of filming. This all commenced as soon as the cast was established, prior to read-through in large part.

Rehearsal and Editing Timeline:

Six weeks were allotted between the initiation of rehearsal and show premiere. The first half of the show was rehearsed for two weeks before filming off-book took place for the third week. Rehearsals for the second half began that same week, concurrent to pickup filming sessions to make up for inadequate footage or scheduling conflicts for actors. Pickups were very challenging to reliably schedule, and should be avoided as much as possible. Filming and final pickups for the second half took place over the fifth week. Rough scene editing, as well as editing of inserts like the introduction, montages, and credits all took place concurrently to these stages of filming. Final tight-cutting, music/SFX, color correcting, keying, and assembly took place over the last week by the lead editor. 

Workflow: 

Each editor was tasked with a particular set of individual scenes to conduct a “rough cut”, or select best takes and align logically as outlined by the script. Editors were expected to read their scenes, as well as attend at least one rehearsal along with the recording session. Timing did not always allow for this, but thorough notes were taken on behalf of relevant editors by (A)SMs and any other editors present (generally the lead). Some sessions were recorded via Zoom for later viewing by editors in order to capture specific director commentary. In future, this should be standardized for all sessions with editors absent. The lead editor was present for all filming sessions to guide filming and ensure checklists were being followed [https://docs.google.com/document/d/1ZNAnlqo7ZFSt8V7jom4Cr0tQyMenJrrDvShgs_G2ZU8/edit], as well as determine if the desired footage was achieved. This meant setting the energy for a scene alongside the director, and keeping a close eye on eye-lines, frame edges, blocking, etc. After ‘cut’ on a take was called, the lead would elect for another take in the same camera configuration (with editor and director feedback as needed), or reset for a different camera angle.

Footage and audio from actors were uploaded to a bespoke shared Dropbox folder, subdivided by scene, and further subdivided by character. File naming conventions were established (but not always followed; some renaming had to occur, or clarification with actors where unclear). Some special accommodations were established for actors with Dropbox sync issues (email/Google Drive uploads), but largely caused no issue. Actors were expected to upload the night of a recording, with the ability to let uploads run overnight in cases of poor internet connection. This mostly worked, but some actors had to be continuously reminded to initiate upload (primarily done by lead editor, and assisted by SMs). Some sort of accountability system should be in-built in future - dedicated time in rehearsals to ensure that uploads are started is one option that has proven fairly successful, but eats into scheduled rehearsal time pretty considerably. 

Rough cuts were expected approximately one week after completion of a scene. This is very achievable for standard-length scenes, and was primarily bottlenecked by a lack of footage due to actor scheduling or lack of upload. Rough cuts varied in difficulty with the number of characters in a scene. Rough cut editors for a scene were expected to watch all footage associated with a scene, working initially from best overall takes as well as splicing in the best moments from alternate takes to create the most cohesive sequence from available footage. This splicing could be more or less meticulous (editor-dependent), and often involved the lead editor later swapping in selects from alternate takes while working through the pacing-focused pass of a scene: the “tight cut.”

Tight cuts were an extremely time consuming stage in this process, and required thorough understanding of dynamics within a scene, as well as consideration of the scene in its place within the play as a whole. For this project, pacing was best understood chronologically within a scene, being extremely conscious of dead time as opposed to conscious beats in conversation. Relying on the timing from the live recordings is not adequate due to 1) the inclusion of elements from different takes or video inserts; 2) timing challenges, connectivity interruption, inability to hear precise cues, or momentary resets while filming was still rolling; and 3) incorporation of video and/or audio reactions within a conversation. For example, a moment may play faster in recording and require extending a beat for a reaction to read properly. In opposition to this was the most important part of this stage: simply “tightening.” Momentum and energy from beginning to end of a scene had to flow in a logical and engaging manner; the most common and tedious element here was therefore adjusting the timing of conversation (dialogue and reaction) to fit the energetic arc necessary for the scene, and feel as “natural” as possible. Seriously, this takes forever, and rushing this step will result in a scene that feels clunky/awkward.

For high energy, quick cuts (about 3-5 seconds each) were used liberally via the introduction of digital zooms, different angles of the same speaker, and reaction shots for scene partners. For lower, more somber or thoughtful energy, longer cuts, more silent spells, and general reliance on ambience and dwelling on natural actor pauses were leveraged.

This tight cut step therefore included the introduction of sound effects (acute and ambient), and subsequently music (either original compositions or royalty free background tracks, usually downloaded from YouTube). A final pass for timing is critical relative to music shifts etc. At this stage, director feedback is extremely helpful to ensure that the correct feeling for a scene is evoked.

Finally, visual elements were introduced, including color correction, keying out green screens, and adding in overlays like title cards or lower thirds. 

Rendering and file sharing:

Editors ran into some inconsistencies using the Team Project capability of Premiere Pro, and therefore instead shared full sequence and data folders for each scene on Google Drive. Video was rendered and deposited in a Drive folder for each scene for director feedback (usually at the “tight cut” phase, with feedback incorporation denoted by version/date and LOTS of communication directly with the director). A document for feedback was maintained alongside these different versions in Drive as a living document. [link] To simplify the final sequence assembly process and avoid accidental changes/retiming, individual scenes were rendered and assembled as such with the incorporation of transitions and other video inserts. Pacing of transitions also matters (!!), so once the pacing within a scene was concretized, the flow between scenes was necessary to fit into the context created by its surroundings. This sometimes meant leveraging jarring sound effects, cutting to black, fading sound in first, cutting directly between montages, etc.

This destructive workflow technique (rendering pieces to re-render the whole) was mostly undertaken for the sanity of the lead editor, and may have resulted in a few minor encoding issues. Still, this was a successful and effective choice to cut down on final rendering time, reduce mishaps by condensing the number of interacting elements in a video sequence, and create timing blocks that could be easily swapped out with newer versions of scenes as needed while including version information by scene. 

TL;DR and Main Takeaways:

• Workflow from filming, to re-shoots or pick-ups, rough editing (selects), tight editing (pacing, SFX/music), and visual editing (VFX/color) should be tracked, and ownership by editors of different steps in a scene clearly defined.
• Scenes (treated discretely in the above workflow) should be laid into the final sequence with conscious planning for flow in transition.
• Avoid pick-up shoots at all costs!!! The biggest bottleneck is getting all the necessary footage for a scene to the editors.
• This project involved the lead editor acting as both cinematographer/DP and primary editor. This meant that a lot of time was sunk into guiding/attending filming sessions, therefore pushing work on actual editing off until filming was largely concluded. These should be two different people (though they should work closely), with greater reliance on notes from (A)SMs in a scene for editors to gather necessary information.
• A strategically destructive workflow is a life-saver. Don’t over-/under-do, or you will go crazy within the timeframe necessary for a production.
• The most time-consuming and feedback-critical phase is the tight cut. Make sure a shared vocabulary and feedback workflow is established with the director to make meaningful changes. Expectations about the degree of specificity that is useful should be established in advance.
• Unresolved: How to create a cohesive “voice” without having the entire tight-cut stage rest on the lead editor? Possibly establishing greater expectation on the refinement of rough cuts will assist with this.

DoP Best Practices/Notae Bene:

• Scenes were usually constructed with the inclusion of two to three stationary camera points, around which angles to sight-lines etc. were constructed. For the sake of time, this was not rigidly followed, but did provide for a good baseline in preparing for shooting a scene.
• Better to over-light a shot than under-light (especially be aware of adequately lighting the green screen, if in use) - if information is lost, keying and recoloring becomes extremely challenging; smartphone cameras generally handle more light better by default.
• Be extremely sure that a camera is level with the ground!! It is very easy to not double check this, and though it can be corrected in post, is a bit of a headache for framing if extremely off-kilter.
• Be wary of reflective things (glasses, tape holding up the background, jewelry - it all can compromise keying).
• If possible, especially for dynamic shots where color consistency is important, manually lock the exposure of the smartphone (else it will try to compensate and result in dramatic color shifts).
• Try to build a way to monitor a shot into your set-up - this can mean placing the laptop/Zoom session somewhere behind the filming smartphone so that the director/DoP are able to view the filming device’s screen through the Zoom device’s camera; factoring in a period of remote framing set-up via Zoom as well was helpful, as actors could log into the Zoom session via their filming device before filming started for the director/DoP to provide notes/tweaks (note that most smartphones will have a variably zoomed in frame for the on-device filming application vs the image that appears in the Zoom session; the lighting and center are maintained, and degree of zooming can be estimated).
• Height matters! Proximity matters! Factor these in and how dynamics due to blocking or actor choices will affect framing. Always prefer more space (ie a wider shot) and a digital zoom in post if lighting and resolution are not a concern.