How to Edit Hours of Interview Footage Without Losing Your Mind

The most efficient way to edit large volumes of interview footage is to separate watching from deciding. Watch everything once with a scoring system, build a ranked selects pool from the top-scored moments, then assemble your story from that curated pool. Trying to watch, evaluate, and build simultaneously is what turns a 2-day edit into a 2-week ordeal.

I learned this the hard way on my third documentary project, sitting in front of 87 hours of raw interview footage across 22 subjects. I had no system. I just started watching and cutting. Three weeks later, I had a rough cut that did not work because I had lost perspective on the material. The moments I selected first (when I was fresh and engaged) were strong. The moments I selected on day eight (when I was exhausted and just wanted to finish) were not.

The system I developed after that experience has guided every project since. Here is how it works.

The Volume Problem Is a Cognitive Problem

The challenge with large-volume interview footage is not technical. Your NLE can handle hundreds of hours of media. The challenge is cognitive. Human attention and editorial judgment degrade over time. By hour six of continuous footage review, you are not making the same quality decisions you were making in hour one.

This is why the most common editing mistake on interview-heavy projects is not bad selection. It is inconsistent selection. The early interviews get careful attention. The later interviews get scanned. The result is a rough cut where the first act is strong and the third act is weak, not because the footage is weaker, but because the editor's attention was weaker.

Research on decision fatigue confirms what editors experience intuitively. The quality of decisions degrades with volume. The solution is structure: systems that separate the high-attention task (evaluating quality) from the low-attention task (mechanical assembly).

Step 1: Chunk Your Footage Into Sessions

Do not attempt to watch everything in one sitting. Break your footage into review sessions of no more than 90 minutes each. Schedule these sessions across multiple days if possible.

For a project with 6 hours of raw footage across 4 interview subjects, that might look like: Day 1 morning, review Subject A (90 minutes). Day 1 afternoon, review Subject B (90 minutes). Day 2 morning, review Subject C (90 minutes). Day 2 afternoon, review Subject D (90 minutes). Day 3, build paper edit and pull selects.

The gap between sessions is not wasted time. Your brain continues processing the material unconsciously. Editors who sleep on footage before making final selections consistently produce better work than editors who power through everything in one marathon session.

Step 2: Score as You Watch

Develop a simple scoring system and apply it consistently across all footage. My system uses three tiers.

Tier 1 (must use): Moments where the subject says something essential to the story with strong delivery. These are the foundational clips that the rough cut cannot exist without.

Tier 2 (strong candidate): Moments with good content or good delivery, but not both at the highest level. These are the clips you draw from when building transitions, adding depth, or replacing Tier 1 clips that do not work in context.

Tier 3 (notable but not essential): Interesting tangents, humanizing asides, potential B-roll voice-over candidates. You probably will not use these in the main cut, but they are worth marking for potential use.

Mark each moment in your NLE with a color-coded marker or label as you watch. Do not stop to trim or evaluate deeply. Just score and move on. The goal is to get through all footage with consistent attention quality.

At the end of this phase, you have a scored map of every interview. Tier 1 moments are your core material. Tier 2 fills gaps. Tier 3 is your backup inventory.

Step 3: Build a Selects Pool by Tier

After all footage is scored, pull Tier 1 moments into a dedicated selects sequence. Watch them back-to-back without the surrounding footage. This is often revelatory. Moments that felt powerful in context sometimes feel flat in isolation, and vice versa. Adjust your scoring if needed.

Then pull Tier 2 moments into a separate sequence. You now have two organized pools to draw from during assembly.

This tiered approach means you never need to re-watch raw footage during the assembly phase. Everything you need is in one of two sequences, already evaluated and scored.

Step 4: Assemble From Selects, Not From Raw

Build your rough cut exclusively from the selects pools. Resist the temptation to "just check" the raw footage for something you might have missed. If you scored consistently, you did not miss anything important. Returning to raw footage during assembly is the number one cause of scope creep and schedule overruns.

Start with Tier 1 selects. Arrange them in narrative order following your paper edit. Fill structural gaps with Tier 2 selects. If a section feels thin, it usually means the subject did not provide strong material on that topic, and the right solution is to restructure the narrative rather than dig through raw footage hoping to find something you missed.

Step 5: Take a Full Day Off Before the Pacing Pass

If your schedule allows it, put one full day between the rough assembly and the pacing pass. When you return, watch the rough cut with fresh eyes. You will immediately see structural problems and pacing issues that were invisible when you were deep in the material.

This distance is not a luxury. It is a technique. Professional editors who build "cool down" time into their schedules consistently deliver better work on the same budget as editors who work straight through.

How AI Changes the Equation

The system I just described is the manual approach, and it works reliably across any project size. But it is time-intensive. For a project with 6 hours of raw footage, the full process takes 3-5 working days.

AI rough cut tools attack the most time-consuming phases: the initial review, scoring, and selects building. Tools that use prosodic analysis (like Threadline Studio) effectively perform the scoring step automatically, identifying high-delivery moments across all footage and assembling them into a structured narrative. The editor receives a pre-built rough cut as an XML file and starts at the refinement phase rather than the review phase.

This does not eliminate the need for editorial judgment. You still evaluate the AI's selections, restructure as needed, and apply your creative vision. But it compresses a 3-5 day process into a 1-2 day process by handling the mechanical phases that degrade with volume and fatigue.

Frequently Asked Questions

How do you stay focused when editing hours of footage? Break footage into 90-minute review sessions, score rather than cut, and take breaks between sessions. Decision quality degrades after 2+ hours of continuous footage review.

What is the best way to organize interview footage? Use a bin-per-subject structure with color-coded markers for quality tiers. Separate Tier 1 (must use), Tier 2 (strong candidate), and Tier 3 (notable) moments into dedicated selects sequences.

How long does it take to edit a documentary with 20+ hours of footage? Manually, a documentary with 20+ hours of raw interviews typically requires 4-8 weeks of post-production. AI-assisted workflows can compress the initial selects and assembly phases to 1-2 weeks, leaving more time for creative refinement.

Should I watch all the footage before I start editing? Yes. Editors who start cutting before reviewing all material consistently produce weaker work because they over-invest in early footage and under-invest in later footage. Score everything first, then assemble.

Can AI help with large-volume interview editing? Yes. AI tools excel at processing large volumes consistently. Unlike human editors, AI does not experience decision fatigue and evaluates every moment with the same analytical rigor, whether it is minute 5 or minute 500.