The case for cutting on the original timeline

The offline/online split is a historical artifact. It exists because thirty years ago broadcast-quality images couldn't sit on the editor's workstation, and the colorist's bay was a different room with different hardware. Neither condition holds on most of the work we do now. A Resolve license and a calibrated reference monitor cover the picture path from selects through delivery, on one timeline.

We still leave the original timeline on some jobs, and we'll say which. But on a small festival feature, on a music documentation cut, on a sound-led commission for a label — we've stopped pretending the round-trip is free.

What the split was actually solving for

Offline editing as a discrete stage was a response to the cost of broadcast-quality VTRs and central storage. The editor cut a low-resolution copy. The online editor reassembled the show at full resolution from an EDL or AAF and the camera-original tapes. The conform existed because the offline NLE could not render broadcast-quality images. That was true in 1989, when Avid Media Composer first shipped. It isn't true any more.

DNxHD and DNxHR gave Avid editors lightweight proxies that could be swapped back to the camera original at full resolution without leaving the project. ProRes Proxy did the same for Final Cut Pro. Premiere added optimized media and proxy media. Resolve unified edit, color, Fusion compositing, and Fairlight audio on a single timeline. As Filmmaker Magazine has observed, the line between offline-compressed and online-full has been blurring for a decade — the codec problem is solved.

What that means in practice: the picture editor working in Resolve can switch between a DNxHR proxy and the camera-original log file inside the same project, color against the camera original, and deliver a DCP-grade master without ever opening a second application. The technical reason for a separate online stage has gone away on productions that don't need a finishing house.

When the round-trip is overhead

The conform is where entropy enters. Each XML or AAF hand-off is a place for the wrong frame rate to land, the wrong source media to be linked, an effect to render against a proxy aspect ratio. Each conform demands verification time the schedule rarely accounts for. Wikipedia's account of the offline edit describes the offline/online split as a workflow that "remained dominant for the next 20 years" — and those twenty years ended a while ago.

The cascade after picture lock is worse. As composers have argued for years, every frame change after lock taxes the music cues, the foley, the ADR, and the M&E stem. When a feature picture-locks on one timeline and then reconforms in a separate online session, every department downstream of picture has to verify that the conform matches the EDL or AAF that went out at lock. Fewer reconforms between picture lock and final delivery, safer audio side of the calendar.

On a one-room production, where the picture editor is also the colorist — increasingly true on the work we do — the round-trip is pure overhead. We're not handing the cut off. We're handing it to ourselves, in the same machine, on the same timeline, with an extra opportunity for a frame to slip.

When the round-trip earns its keep

This isn't an argument that the offline/online split is always wrong. VFX-heavy work needs the discrete handoff. Multi-format deliverables — theatrical DCP plus festival H.264 plus broadcast versions — usually want a finishing pass organized around delivery, not creative editing. Productions where the colorist is a different person in a different city want the EDL and the audit trail.

IndieWire's post-production pitfalls list is not a counsel against ever leaving the original timeline. It's a counsel against treating the round-trip as free. On a feature where one operator owns picture from selects through master, the trip is rarely free. On a feature with discrete dailies grading and a colorist who needs a different bay, it is. The decision belongs to the production, not to the workflow diagram on the wall.

What it looks like to deliver from the original timeline

When we assemble a cut here, we start in the timeline we intend to deliver from. Proxy switching lives in the same project. Color happens against the camera original in the same sequence. The picture lock we send to the composer is the picture lock that makes the master. There is one EDL in the project's history — the one nobody had to send. Sound, the cascade we most want to protect, sits at the end of a shorter chain, and the chain has fewer joints to fail at.