Turns one take per question into an editable interview film. The speaker is cut out onto brand backgrounds at a consistent size and cut together with question cards, with square and widescreen social versions.
Presentation & Video Production
Cross-sector
Pipeline
Productised
Local/private
2026
The problem
Interview answers are shot in different rooms at different head sizes, then edited by hand onto slides. Square and widescreen social cuts become a second project that drifts from the master.
What it does
A person supplies one clip per question, the brand plates and the question list. Each answer is cut off its real background, framed to one head size, trimmed to the speech and placed on a brand plate with a question card. The sequence stays editable in Premiere, and the square and landscape social cuts are exports of the same questions.
Outcomes
Answers from different rooms share one head size and background.
The editor still has an editable sequence.
Social versions match the questions in the long film.
The start and end of speech are measured and recorded in a report.
Simplified mockup
Interactive preview with sample data. Nothing is sent to a server.
One take per question
Takes
3
Paired to the list
Spoken length
2:25
Trimmed to speech
Head size
18%
One target
Those counts are the planned report. This sample does not measure a filmed speaker.
Bramble Works · site interview Rowan Hale
01
What changed on site this year?
q1-this-year.mov
00:42
02
Why not just hire someone?
q2-hire-or-build.mov
01:05
03
Where should a small firm start?
q3-start-here.mov
00:38
Assembly
Three takes waiting
Editable sequence Not a flatten
bramble-works-interview.xml
The Player tab plays the sample film. A filmed speaker, matte and head-size pass are not in this file.
CardQ1
AnswerPlate 01
CardQ2
AnswerPlate 02
CardQ3
AnswerPlate 03
Sample film
No filmed speaker
Intro and outro stings, question cards, lower thirds and captions over engraved plates rendered in Blender. Stock narration. Chapter buttons seek this file. Square crops the same picture.
Framing and speech
Sample figures. Nothing in this table was measured from a filmed take.
Mean head height
18.2%
Spread
0.4
points across three rooms
Bounds
Measured
Not guessed
Take
File
Head
Speech in
Speech out
Duration
01
q1-this-year.mov
18.2%
00:01.4
00:40.9
00:42
02
q2-hire-or-build.mov
18.0%
00:00.6
01:03.8
01:05
03
q3-start-here.mov
18.4%
00:02.1
00:36.4
00:38
Same answer, two crops
Diagram of the crop only. The Player tab has the narrated film. There is no filmed speaker in this sample, so the head box is not a measurement.
16:9 masterWhat changed on site this year?Rowan Hale
Widescreen · question beside the plate
1:1What changed on site this year?Q1 · What changed on site?
Rowan HaleOperations lead
Square head crop
How it works
A folder of answers is paired with the question list.
The speaker is cut out from the real room onto a brand background.
Framing is evened out to a consistent head size and each clip is trimmed to the speech.
Question cards are placed between answers in an editable Premiere sequence.
A widescreen film is assembled, and square and landscape versions are exported for each question.
A short report records the framing and where speech starts and ends.
Pipeline
flowchart LR
A[Answer clips] --> B[Matte]
B --> C[Head size]
C --> D[Trim to speech]
E[Question cards] --> F[Sequence]
D --> F
F --> G[Widescreen film]
F --> H[Social cuts]
Output
Poster from the sample interview film
Sample output
Intro and outro sting, question cards, lower thirds and captions over engraved plates rendered in Blender, with stock narration. No filmed speaker.